From d0c12ec952b03607be05e563c7ea5fe0ee307dcf Mon Sep 17 00:00:00 2001
From: johannes wasmer <johannes.wasmer@gmail.com>
Date: Tue, 20 Aug 2024 13:40:09 +0100
Subject: [PATCH] bibliography update
---
bib/bibliography.bib | 3488 +++++++++++++++++++++++++++++++-----------
1 file changed, 2617 insertions(+), 871 deletions(-)
diff --git a/bib/bibliography.bib b/bib/bibliography.bib
index 69ae34d..8b8d57b 100644
--- a/bib/bibliography.bib
+++ b/bib/bibliography.bib
@@ -6,7 +6,7 @@
abstract = {The discovery of the quantum Hall effect in 1980 marked a turning point in condensed matter physics.},
langid = {english},
organization = {Nature},
- keywords = {/unread,2D material,ARPES,Berry phase,collection,Hall effect,Hall QHE,heterostructures,Heusler,history of science,magnetic order,magnetism,popular science,quantum materials,semimetal,strongly correlated maeterials,superconductor,TMDC,topological,topological insulator,topological phase,vdW materials,Weyl semimetal},
+ keywords = {2D material,ARPES,Berry phase,collection,Hall effect,Hall QHE,heterostructures,Heusler,history of science,magnetic order,magnetism,popular science,quantum materials,semimetal,strongly correlated maeterials,superconductor,TMDC,topological,topological insulator,topological phase,vdW materials,Weyl semimetal},
file = {/Users/wasmer/Zotero/storage/SJH8NYEP/fdbjbijfea.html}
}
@@ -23,6 +23,45 @@
file = {/Users/wasmer/Zotero/storage/GX4WMQWH/Surrogate models for the electron density and related scalar fields.html}
}
+@article{abrahamFusingMachineLearning2023,
+ title = {Fusing a Machine Learning Strategy with Density Functional Theory to Hasten the Discovery of {{2D MXene-based}} Catalysts for Hydrogen Generation},
+ author = {Abraham, B. Moses and Sinha, Priyanka and Halder, Prosun and Singh, Jayant K.},
+ date = {2023-04-11},
+ journaltitle = {Journal of Materials Chemistry A},
+ shortjournal = {J. Mater. Chem. A},
+ volume = {11},
+ number = {15},
+ pages = {8091--8100},
+ publisher = {The Royal Society of Chemistry},
+ issn = {2050-7496},
+ doi = {10.1039/D3TA00344B},
+ url = {https://pubs.rsc.org/en/content/articlelanding/2023/ta/d3ta00344b},
+ urldate = {2024-05-28},
+ abstract = {The complexity of the topological and combinatorial configuration space of MXenes can give rise to gigantic design challenges that cannot be addressed through traditional experimental or routine theoretical methods. To this end, we establish a robust and more broadly applicable multistep workflow using supervised machine learning (ML) algorithms to construct well-trained data-driven models for predicting the hydrogen evolution reaction (HER) activity of 4500 MM′XT2-type MXenes, where 25\% of the materials space (1125 systems) is randomly selected to evaluate the HER performance using density functional theory (DFT) calculations. As the most desirable ML model, the gradient boosting regressor (GBR) processed with recursive feature elimination (RFE), hyperparameter optimization (HO) and the leave-one-out (LOO) approach accurately and rapidly predicts the Gibbs free energy of hydrogen adsorption (ΔGH) with a low predictive mean absolute error (MAE) of 0.358 eV. Based on these observations, the H atoms adsorbed directly on top of the outermost metal-atom layer of the MM′XT2-type MXenes (site 1) with Nb, Mo and Cr metals with O functionalization are discovered to be highly stable and active for catalysis, surpassing commercially available platinum-based counterparts. Overall, the physically meaningful predictions and insights of the developed ML/DFT-based multistep workflow will open new avenues for accelerated screening, rational design and discovery of potential HER catalysts.},
+ langid = {english},
+ keywords = {/unread,AML,catalysis,chemistry,compositional descriptors,cross-validation,gradient boosting,hydrogen evolution reaction,ML,property prediction,random forest,scikit-learn,with-code,workflows},
+ file = {/Users/wasmer/Nextcloud/Zotero/Abraham et al_2023_Fusing a machine learning strategy with density functional theory to hasten the.pdf;/Users/wasmer/Zotero/storage/KVTRJAFE/Abraham et al. - 2023 - Fusing a machine learning strategy with density fu.pdf}
+}
+
+@article{abramsonAccurateStructurePrediction2024,
+ title = {Accurate Structure Prediction of Biomolecular Interactions with {{AlphaFold}} 3},
+ author = {Abramson, Josh and Adler, Jonas and Dunger, Jack and Evans, Richard and Green, Tim and Pritzel, Alexander and Ronneberger, Olaf and Willmore, Lindsay and Ballard, Andrew J. and Bambrick, Joshua and Bodenstein, Sebastian W. and Evans, David A. and Hung, Chia-Chun and O’Neill, Michael and Reiman, David and Tunyasuvunakool, Kathryn and Wu, Zachary and ŽemgulytÄ—, AkvilÄ— and Arvaniti, Eirini and Beattie, Charles and Bertolli, Ottavia and Bridgland, Alex and Cherepanov, Alexey and Congreve, Miles and Cowen-Rivers, Alexander I. and Cowie, Andrew and Figurnov, Michael and Fuchs, Fabian B. and Gladman, Hannah and Jain, Rishub and Khan, Yousuf A. and Low, Caroline M. R. and Perlin, Kuba and Potapenko, Anna and Savy, Pascal and Singh, Sukhdeep and Stecula, Adrian and Thillaisundaram, Ashok and Tong, Catherine and Yakneen, Sergei and Zhong, Ellen D. and Zielinski, Michal and ŽÃdek, Augustin and Bapst, Victor and Kohli, Pushmeet and Jaderberg, Max and Hassabis, Demis and Jumper, John M.},
+ date = {2024-06},
+ journaltitle = {Nature},
+ volume = {630},
+ number = {8016},
+ pages = {493--500},
+ publisher = {Nature Publishing Group},
+ issn = {1476-4687},
+ doi = {10.1038/s41586-024-07487-w},
+ url = {https://www.nature.com/articles/s41586-024-07487-w},
+ urldate = {2024-08-03},
+ abstract = {The introduction of AlphaFold\,21 has spurred a revolution in modelling the structure of proteins and their interactions, enabling a huge range of applications in protein modelling and design2–6. Here we describe our AlphaFold\,3 model with a substantially updated diffusion-based architecture that is capable of predicting the joint structure of complexes including proteins, nucleic acids, small molecules, ions and modified residues. The new AlphaFold model demonstrates substantially improved accuracy over many previous specialized tools: far greater accuracy for protein–ligand interactions compared with state-of-the-art docking tools, much higher accuracy for protein–nucleic acid interactions compared with nucleic-acid-specific predictors and substantially higher antibody–antigen prediction accuracy compared with AlphaFold-Multimer v.2.37,8. Together, these results show that high-accuracy modelling across biomolecular space is possible within a single unified deep-learning framework.},
+ langid = {english},
+ keywords = {/unread,AI4Science,AlphaFold,biomolecules,DeepMind,generative models,Protein structure predictions,structure prediction,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Abramson et al_2024_Accurate structure prediction of biomolecular interactions with AlphaFold 3.pdf}
+}
+
@article{acharMachineLearningElectron2023,
title = {Machine {{Learning Electron Density Prediction Using Weighted Smooth Overlap}} of {{Atomic Positions}}},
author = {Achar, Siddarth K. and Bernasconi, Leonardo and Johnson, J. Karl},
@@ -210,13 +249,13 @@
author = {Akiba, Takuya and Sano, Shotaro and Yanase, Toshihiko and Ohta, Takeru and Koyama, Masanori},
date = {2019-07-25},
eprint = {1907.10902},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.1907.10902},
url = {http://arxiv.org/abs/1907.10902},
urldate = {2023-11-17},
abstract = {The purpose of this study is to introduce new design-criteria for next-generation hyperparameter optimization software. The criteria we propose include (1) define-by-run API that allows users to construct the parameter search space dynamically, (2) efficient implementation of both searching and pruning strategies, and (3) easy-to-setup, versatile architecture that can be deployed for various purposes, ranging from scalable distributed computing to light-weight experiment conducted via interactive interface. In order to prove our point, we will introduce Optuna, an optimization software which is a culmination of our effort in the development of a next generation optimization software. As an optimization software designed with define-by-run principle, Optuna is particularly the first of its kind. We will present the design-techniques that became necessary in the development of the software that meets the above criteria, and demonstrate the power of our new design through experimental results and real world applications. Our software is available under the MIT license (https://github.com/pfnet/optuna/).},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Bayesian methods,Bayesian optimization,black-box optimization,distributed computing,General ML,hyperparameters,hyperparameters optimization,library,ML,Optuna,original publication,PyTorch,scikit-learn,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Akiba et al_2019_Optuna.pdf;/Users/wasmer/Zotero/storage/5LHKTHSE/1907.html}
}
@@ -340,7 +379,7 @@
volume = {8},
number = {1},
eprint = {2103.02068},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
pages = {217},
issn = {2052-4463},
doi = {10.1038/s41597-021-00974-z},
@@ -357,13 +396,13 @@
author = {Anderson, Brandon and Hy, Truong-Son and Kondor, Risi},
date = {2019-11-25},
eprint = {1906.04015},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.1906.04015},
url = {http://arxiv.org/abs/1906.04015},
urldate = {2022-10-04},
abstract = {We propose Cormorant, a rotationally covariant neural network architecture for learning the behavior and properties of complex many-body physical systems. We apply these networks to molecular systems with two goals: learning atomic potential energy surfaces for use in Molecular Dynamics simulations, and learning ground state properties of molecules calculated by Density Functional Theory. Some of the key features of our network are that (a) each neuron explicitly corresponds to a subset of atoms; (b) the activation of each neuron is covariant to rotations, ensuring that overall the network is fully rotationally invariant. Furthermore, the non-linearity in our network is based upon tensor products and the Clebsch-Gordan decomposition, allowing the network to operate entirely in Fourier space. Cormorant significantly outperforms competing algorithms in learning molecular Potential Energy Surfaces from conformational geometries in the MD-17 dataset, and is competitive with other methods at learning geometric, energetic, electronic, and thermodynamic properties of molecules on the GDB-9 dataset.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {equivariant,GNN,MD17,ML,MLP,MPNN,O(3),QM9,representation learning,SchNet,SO(3)},
file = {/Users/wasmer/Nextcloud/Zotero/Anderson et al_2019_Cormorant.pdf;/Users/wasmer/Zotero/storage/RY359LWP/1906.html}
}
@@ -390,7 +429,7 @@
}
@article{andraeHypothesesPrimaryEnergy2020,
- title = {Hypotheses for {{Primary Energy Use}}, {{Electricity Use}} and {{CΟ2 Emissions}} of {{Global Computing}} and {{Its Shares}} of the {{Total Between}} 2020 and 2030},
+ title = {Hypotheses for {{Primary Energy Use}}, {{Electricity Use}} and {{CO2 Emissions}} of {{Global Computing}} and {{Its Shares}} of the {{Total Between}} 2020 and 2030},
author = {Andrae, Anders S. G.},
date = {2020},
journaltitle = {WSEAS Transactions on Power Systems},
@@ -402,8 +441,8 @@
urldate = {2023-08-30},
abstract = {There is no doubt that the economic and computing activity related to the digital sector will ramp up faster in the present decade than in the last. Moreover, computing infrastructure is one of three major drivers of new electricity use alongsidefuture and current hydrogen production and battery electric vehicles charging. Here is proposed a trajectory in this decade for CO2 emissions associated with this digitalization and its share of electricity and energy generation as a whole. The roadmap for major sources of primary energy and electricity and associated CO2 emissions areprojected and connected to the probable power use of the digital industry. The truncation error for manufacturing related CO2 emissions may be 0.8 Gt or more indicating a larger share of manufacturing and absolute digital CO2 emissions.While remaining at a moderate share of global CO2 emissions (4-5\%), the resulting digital CO2 emissions will likely rise from 2020 to 2030. The opposite may only happen if the electricity used to run especially data centers and production plants is produced locally (next to the data centers and plants) from renewable sources and data intensity metrics grow slower than expected.},
langid = {english},
- keywords = {/unread,ecological footprint,economics,energy consumption,energy efficiency,environmental impact,for introductions,ICT sector,low-power electronics,world energy consumption},
- file = {/Users/wasmer/Nextcloud/Zotero/Andrae_2020_Hypotheses for Primary Energy Use, Electricity Use and CΟ2 Emissions of Global.pdf}
+ keywords = {ecological footprint,economics,energy consumption,energy efficiency,environmental impact,for introductions,ICT sector,low-power electronics,world energy consumption},
+ file = {/Users/wasmer/Nextcloud/Zotero/Andrae_2020_Hypotheses for Primary Energy Use, Electricity Use and CO2 Emissions of Global.pdf}
}
@article{andrejevicMachineLearningSpectralIndicators2022,
@@ -429,13 +468,13 @@
author = {Angelopoulos, Anastasios N. and Bates, Stephen and Fannjiang, Clara and Jordan, Michael I. and Zrnic, Tijana},
date = {2023-02-16},
eprint = {2301.09633},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, q-bio, stat},
doi = {10.48550/arXiv.2301.09633},
url = {http://arxiv.org/abs/2301.09633},
urldate = {2023-03-01},
abstract = {We introduce prediction-powered inference \$\textbackslash unicode\{x2013\}\$ a framework for performing valid statistical inference when an experimental data set is supplemented with predictions from a machine-learning system. Our framework yields provably valid conclusions without making any assumptions on the machine-learning algorithm that supplies the predictions. Higher accuracy of the predictions translates to smaller confidence intervals, permitting more powerful inference. Prediction-powered inference yields simple algorithms for computing valid confidence intervals for statistical objects such as means, quantiles, and linear and logistic regression coefficients. We demonstrate the benefits of prediction-powered inference with data sets from proteomics, genomics, electronic voting, remote sensing, census analysis, and ecology.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AlphaFold,DeepMind,General ML,ML,RCPS,risk,uncertainty quantification,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Angelopoulos et al_2023_Prediction-Powered Inference.pdf;/Users/wasmer/Zotero/storage/VUQUZZ32/2301.html}
}
@@ -497,14 +536,14 @@
author = {Antunes, Luis M. and Butler, Keith T. and Grau-Crespo, Ricardo},
date = {2023-07-10},
eprint = {2307.04340},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2307.04340},
url = {http://arxiv.org/abs/2307.04340},
urldate = {2023-07-12},
abstract = {The generation of plausible crystal structures is often an important step in the computational prediction of crystal structures from composition. Here, we introduce a methodology for crystal structure generation involving autoregressive large language modeling of the Crystallographic Information File (CIF) format. Our model, CrystaLLM, is trained on a comprehensive dataset of millions of CIF files, and is capable of reliably generating correct CIF syntax and plausible crystal structures for many classes of inorganic compounds. Moreover, we provide general and open access to the model by deploying it as a web application, available to anyone over the internet. Our results indicate that the model promises to be a reliable and efficient tool for both crystallography and materials informatics.},
- pubstate = {preprint},
- keywords = {AML,autoregressive,CIF,crystal structure,LLM,materials,ML,NOMAD,OQMD,PBE,prediction of structure,pretrained models,transformer,VASP,with-code,with-demo},
+ pubstate = {prepublished},
+ keywords = {AML,autoregressive,CIF,crystal structure,language models,LLM,materials,ML,NOMAD,OQMD,PBE,prediction of structure,pretrained models,transformer,VASP,with-code,with-demo},
file = {/Users/wasmer/Nextcloud/Zotero/Antunes et al_2023_Crystal Structure Generation with Autoregressive Large Language Modeling.pdf;/Users/wasmer/Zotero/storage/VMYV5IJR/2307.html}
}
@@ -670,13 +709,30 @@ Subject\_term: Condensed-matter physics, Materials science},
abstract = {Geometric deep learning (GDL) is based on neural network architectures that incorporate and process symmetry information. GDL bears promise for molecular modelling applications that rely on molecular representations with different symmetry properties and levels of abstraction. This Review provides a structured and harmonized overview of molecular GDL, highlighting its applications in drug discovery, chemical synthesis prediction and quantum chemistry. It contains an introduction to the principles of GDL, as well as relevant molecular representations, such as molecular graphs, grids, surfaces and strings, and their respective properties. The current challenges for GDL in the molecular sciences are discussed, and a forecast of future opportunities is attempted. Geometric representations are becoming more important in molecular deep learning as the spatial structure of molecules contains important information about their properties. Kenneth Atz and colleagues review current progress and challenges in this emerging field of geometric deep learning.},
issue = {12},
langid = {english},
- keywords = {CNN,equivariant,GCN,GDL,GNN,invariance,molecules,MPNN,review,review-of-GDL},
+ keywords = {AML,chemistry,CNN,equivariant,GCN,GDL,GNN,invariance,language models,ML,molecules,MPNN,review,review-of-GDL,RNN,SE(3),symmetry,transformer},
annotation = {Primary\_atype: Reviews\\
Subject\_term: Cheminformatics;Computational models;Computational science\\
Subject\_term\_id: cheminformatics;computational-models;computational-science},
file = {/Users/wasmer/Nextcloud/Zotero/Atz et al_2021_Geometric deep learning on molecular representations.pdf;/Users/wasmer/Zotero/storage/WJWQFR9K/s42256-021-00418-8.html}
}
+@article{b.ghoshClassicalQuantumMachine2023,
+ title = {Classical and Quantum Machine Learning Applications in Spintronics},
+ author = {B.~Ghosh, Kumar J. and Ghosh, Sumit},
+ date = {2023-03-01},
+ journaltitle = {Digital Discovery},
+ volume = {2},
+ number = {2},
+ pages = {512--519},
+ publisher = {Royal Society of Chemistry},
+ doi = {10.1039/D2DD00094F},
+ url = {https://pubs.rsc.org/en/content/articlelanding/2023/dd/d2dd00094f},
+ urldate = {2024-06-05},
+ langid = {english},
+ keywords = {/unread,FZJ,ML,PGI,PGI-1/IAS-1,QML,QSVM,quantum computing,quantum machine learning,quantum transport,random forest,rec-by-ghosh,spin dynamics,spintronics,Spintronics,SVM,tight binding,transport properties},
+ file = {/Users/wasmer/Nextcloud/Zotero/B. Ghosh_Ghosh_2023_Classical and quantum machine learning applications in spintronics.pdf}
+}
+
@article{bacTopologicalResponseAnomalous2022,
title = {Topological Response of the Anomalous {{Hall}} Effect in {{MnBi2Te4}} Due to Magnetic Canting},
author = {Bac, S.-K. and Koller, K. and Lux, F. and Wang, J. and Riney, L. and Borisiak, K. and Powers, W. and Zhukovskyi, M. and Orlova, T. and Dobrowolska, M. and Furdyna, J. K. and Dilley, N. R. and Rokhinson, L. P. and Mokrousov, Y. and McQueeney, R. J. and Heinonen, O. and Liu, X. and Assaf, B. A.},
@@ -723,13 +779,13 @@ Subject\_term\_id: cheminformatics;computational-models;computational-science},
author = {Bakshi, Ainesh and Liu, Allen and Moitra, Ankur and Tang, Ewin},
date = {2023-10-03},
eprint = {2310.02243},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {quant-ph},
doi = {10.48550/arXiv.2310.02243},
url = {http://arxiv.org/abs/2310.02243},
urldate = {2023-10-05},
abstract = {We study the problem of learning a local quantum Hamiltonian \$H\$ given copies of its Gibbs state \$\textbackslash rho = e\textasciicircum\{-\textbackslash beta H\}/\textbackslash textrm\{tr\}(e\textasciicircum\{-\textbackslash beta H\})\$ at a known inverse temperature \$\textbackslash beta{$>$}0\$. Anshu, Arunachalam, Kuwahara, and Soleimanifar (arXiv:2004.07266) gave an algorithm to learn a Hamiltonian on \$n\$ qubits to precision \$\textbackslash epsilon\$ with only polynomially many copies of the Gibbs state, but which takes exponential time. Obtaining a computationally efficient algorithm has been a major open problem [Alhambra'22 (arXiv:2204.08349)], [Anshu, Arunachalam'22 (arXiv:2204.08349)], with prior work only resolving this in the limited cases of high temperature [Haah, Kothari, Tang'21 (arXiv:2108.04842)] or commuting terms [Anshu, Arunachalam, Kuwahara, Soleimanifar'21]. We fully resolve this problem, giving a polynomial time algorithm for learning \$H\$ to precision \$\textbackslash epsilon\$ from polynomially many copies of the Gibbs state at any constant \$\textbackslash beta {$>$} 0\$. Our main technical contribution is a new flat polynomial approximation to the exponential function, and a translation between multi-variate scalar polynomials and nested commutators. This enables us to formulate Hamiltonian learning as a polynomial system. We then show that solving a low-degree sum-of-squares relaxation of this polynomial system suffices to accurately learn the Hamiltonian.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,finite-temperature,ML,ML-QM,ML-QMBP,NQS,prediction of Hamiltonian matrix},
file = {/Users/wasmer/Nextcloud/Zotero/Bakshi et al_2023_Learning quantum Hamiltonians at any temperature in polynomial time.pdf;/Users/wasmer/Zotero/storage/BGGJUKBE/2310.html}
}
@@ -739,13 +795,13 @@ Subject\_term\_id: cheminformatics;computational-models;computational-science},
author = {Balestriero, Randall and Ibrahim, Mark and Sobal, Vlad and Morcos, Ari and Shekhar, Shashank and Goldstein, Tom and Bordes, Florian and Bardes, Adrien and Mialon, Gregoire and Tian, Yuandong and Schwarzschild, Avi and Wilson, Andrew Gordon and Geiping, Jonas and Garrido, Quentin and Fernandez, Pierre and Bar, Amir and Pirsiavash, Hamed and LeCun, Yann and Goldblum, Micah},
date = {2023-04-24},
eprint = {2304.12210},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2304.12210},
url = {http://arxiv.org/abs/2304.12210},
urldate = {2023-05-15},
abstract = {Self-supervised learning, dubbed the dark matter of intelligence, is a promising path to advance machine learning. Yet, much like cooking, training SSL methods is a delicate art with a high barrier to entry. While many components are familiar, successfully training a SSL method involves a dizzying set of choices from the pretext tasks to training hyper-parameters. Our goal is to lower the barrier to entry into SSL research by laying the foundations and latest SSL recipes in the style of a cookbook. We hope to empower the curious researcher to navigate the terrain of methods, understand the role of the various knobs, and gain the know-how required to explore how delicious SSL can be.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {best practices,computer vision,contrastive learning,data augmentation,Deep learning,General ML,guidelines,image data,ML,pretrained models,SimCLR,SSL,transformer,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Balestriero et al_2023_A Cookbook of Self-Supervised Learning.pdf;/Users/wasmer/Zotero/storage/RRIMP57I/2304.html}
}
@@ -774,14 +830,14 @@ Subject\_term\_id: cheminformatics;computational-models;computational-science},
author = {Bao, Ting and Xu, Runzhang and Li, He and Gong, Xiaoxun and Tang, Zechen and Fu, Jingheng and Duan, Wenhui and Xu, Yong},
date = {2024-04-09},
eprint = {2404.06449},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2404.06449},
url = {http://arxiv.org/abs/2404.06449},
urldate = {2024-04-18},
abstract = {Moir\textbackslash 'e-twisted materials have garnered significant research interest due to their distinctive properties and intriguing physics. However, conducting first-principles studies on such materials faces challenges, notably the formidable computational cost associated with simulating ultra-large twisted structures. This obstacle impedes the construction of a twisted materials database crucial for datadriven materials discovery. Here, by using high-throughput calculations and state-of-the-art neural network methods, we construct a Deep-learning Database of density functional theory (DFT) Hamiltonians for Twisted materials named DDHT. The DDHT database comprises trained neural-network models of over a hundred homo-bilayer and hetero-bilayer moir\textbackslash 'e-twisted materials. These models enable accurate prediction of the DFT Hamiltonian for these materials across arbitrary twist angles, with an averaged mean absolute error of approximately 1.0 meV or lower. The database facilitates the exploration of flat bands and correlated materials platforms within ultra-large twisted structures.},
- pubstate = {preprint},
- keywords = {/unread,\_tablet,Condensed Matter - Materials Science},
+ pubstate = {prepublished},
+ keywords = {/unread,Condensed Matter - Materials Science},
file = {/Users/wasmer/Nextcloud/Zotero/Bao et al_2024_Deep-Learning Database of Density Functional Theory Hamiltonians for Twisted.pdf;/Users/wasmer/Zotero/storage/JU4UIG8U/2404.html}
}
@@ -826,17 +882,15 @@ Subject\_term\_id: cheminformatics;computational-models;computational-science},
file = {/Users/wasmer/Nextcloud/Zotero/Barth_Hedin_1972_A local exchange-correlation potential for the spin polarized case.pdf}
}
-@book{bartok-partayGaussianApproximationPotential2010,
+@thesis{bartok-partayGaussianApproximationPotential2010,
title = {The {{Gaussian Approximation Potential}}},
author = {Bartόk-Pártay, Albert},
date = {2010},
- series = {Springer {{Theses}}},
- publisher = {Springer Berlin Heidelberg},
+ institution = {Springer Berlin Heidelberg},
location = {Berlin, Heidelberg},
doi = {10.1007/978-3-642-14067-9},
- url = {http://link.springer.com/10.1007/978-3-642-14067-9},
+ url = {https://doi.org/10.1007/978-3-642-14067-9},
urldate = {2021-07-06},
- isbn = {978-3-642-14066-2 978-3-642-14067-9},
langid = {english},
keywords = {GAP,GPR,ML,models,original publication},
file = {/home/johannes/Books/scientific_machine_learning/Bartók-Pártay_The Gaussian Approximation Potential_thesis-2010.pdf}
@@ -867,12 +921,12 @@ Subject\_term\_id: cheminformatics;computational-models;computational-science},
author = {Bartók, Albert P. and Csányi, Gábor},
date = {2020-02-05},
eprint = {1502.01366},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/1502.01366},
urldate = {2021-07-06},
abstract = {We present a swift walk-through of our recent work that uses machine learning to fit interatomic potentials based on quantum mechanical data. We describe our Gaussian Approximation Potentials (GAP) framework, discussing a variety of descriptors, how to train the model on total energies and derivatives and the simultaneous use of multiple models. We also show a small example using QUIP, the software sandbox implementation of GAP that is available for non-commercial use.},
- keywords = {\_tablet,GAP,ML,models,QUIP,SOAP,tutorial},
+ keywords = {GAP,ML,models,QUIP,SOAP,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Bartók_Csányi_2020_Gaussian Approximation Potentials.pdf;/Users/wasmer/Zotero/storage/SBML3RKM/1502.html}
}
@@ -891,17 +945,34 @@ Subject\_term\_id: cheminformatics;computational-models;computational-science},
file = {/Users/wasmer/Nextcloud/Zotero/Bartók et al_2017_Machine learning unifies the modeling of materials and molecules.pdf;/Users/wasmer/Zotero/storage/DZL84DP7/sciadv.html}
}
-@article{bartokRepresentingChemicalEnvironments2013,
+@online{bartokRepresentingChemicalEnvironments2013,
title = {On Representing Chemical Environments},
- author = {Bartók, Albert P.},
- date = {2013},
+ author = {Bartók, Albert P. and Kondor, Risi and Csányi, Gábor},
+ date = {2013-03-20},
+ doi = {10.48550/arXiv.1209.3140},
+ url = {https://doi.org/10.48550/arXiv.1209.3140},
+ abstract = {We review some recently published methods to represent atomic neighbourhood environments, and analyse their relative merits in terms of their faithfulness and suitability for fitting potential energy surfaces. The crucial properties that such representations (sometimes called descriptors) must have are differentiability with respect to moving the atoms, and invariance to the basic symmetries of physics: rotation, reflection, translation, and permutation of atoms of the same species. We demonstrate that certain widely used descriptors that initially look quite different are specific cases of a general approach, in which a finite set of basis functions with increasing angular wave numbers are used to expand the atomic neighbourhood density function. Using the example system of small clusters, we quantitatively show that this expansion needs to be carried to higher and higher wave numbers as the number of neighbours increases in order to obtain a faithful representation, and that variants of the descriptors converge at very different rates. We also propose an altogether new approach, called Smooth Overlap of Atomic Positions (SOAP), that sidesteps these difficulties by directly defining the similarity between any two neighbourhood environments, and show that it is still closely connected to the invariant descriptors. We test the performance of the various representations by fitting models to the potential energy surface of small silicon clusters and the bulk crystal.},
+ pubstate = {prepublished},
+ keywords = {descriptors,ML,original publication,SOAP},
+ file = {/Users/wasmer/Nextcloud/Zotero/Bartók_2013_On representing chemical environments.pdf;/Users/wasmer/Zotero/storage/VRNA6FAC/PhysRevB.87.html}
+}
+
+@article{bartokRepresentingChemicalEnvironments2013a,
+ title = {On Representing Chemical Environments},
+ author = {Bartók, Albert P. and Kondor, Risi and Csányi, Gábor},
+ date = {2013-05-28},
journaltitle = {Physical Review B},
shortjournal = {Phys. Rev. B},
volume = {87},
number = {18},
+ pages = {184115},
+ publisher = {American Physical Society},
doi = {10.1103/PhysRevB.87.184115},
- keywords = {descriptors,ML,original publication,SOAP},
- file = {/Users/wasmer/Nextcloud/Zotero/Bartók_2013_On representing chemical environments.pdf;/Users/wasmer/Zotero/storage/VRNA6FAC/PhysRevB.87.html}
+ url = {https://link.aps.org/doi/10.1103/PhysRevB.87.184115},
+ urldate = {2024-05-22},
+ abstract = {We review some recently published methods to represent atomic neighborhood environments, and analyze their relative merits in terms of their faithfulness and suitability for fitting potential energy surfaces. The crucial properties that such representations (sometimes called descriptors) must have are differentiability with respect to moving the atoms and invariance to the basic symmetries of physics: rotation, reflection, translation, and permutation of atoms of the same species. We demonstrate that certain widely used descriptors that initially look quite different are specific cases of a general approach, in which a finite set of basis functions with increasing angular wave numbers are used to expand the atomic neighborhood density function. Using the example system of small clusters, we quantitatively show that this expansion needs to be carried to higher and higher wave numbers as the number of neighbors increases in order to obtain a faithful representation, and that variants of the descriptors converge at very different rates. We also propose an altogether different approach, called Smooth Overlap of Atomic Positions, that sidesteps these difficulties by directly defining the similarity between any two neighborhood environments, and show that it is still closely connected to the invariant descriptors. We test the performance of the various representations by fitting models to the potential energy surface of small silicon clusters and the bulk crystal.},
+ keywords = {AML,descriptors,ML,original publication,SB descriptors,SOAP},
+ file = {/Users/wasmer/Nextcloud/Zotero/Bartók et al_2013_On representing chemical environments.pdf;/Users/wasmer/Zotero/storage/NKEUCLSB/PhysRevB.87.html}
}
@article{basovPropertiesDemandQuantum2017,
@@ -945,7 +1016,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
abstract = {We introduce equi-tuning, a novel fine-tuning method that transforms (potentially non-equivariant) pretrained models into group equivariant models while incurring minimum L\_2 loss between the feature representations of the pretrained and the equivariant models. Large pretrained models can be equi-tuned for different groups to satisfy the needs of various downstream tasks. Equi-tuned models benefit from both group equivariance as an inductive bias and semantic priors from pretrained models. We provide applications of equi-tuning on three different tasks: image classification, compositional generalization in language, and fairness in natural language generation (NLG). We also provide a novel group-theoretic definition for fairness in NLG. The effectiveness of this definition is shown by testing it against a standard empirical method of fairness in NLG. We provide experimental results for equi-tuning using a variety of pretrained models: Alexnet, Resnet, VGG, and Densenet for image classification; RNNs, GRUs, and LSTMs for compositional generalization; and GPT2 for fairness in NLG. We test these models on benchmark datasets across all considered tasks to show the generality and effectiveness of the proposed method.},
issue = {6},
langid = {english},
- keywords = {/unread,\_tablet,benchmarking,equivariant,fine-tuning,group theory,image classification,inductive bias,natural language generation,nlp,pretrained models,symmetry},
+ keywords = {/unread,benchmarking,equivariant,fine-tuning,group theory,image classification,inductive bias,natural language generation,nlp,pretrained models,symmetry},
file = {/Users/wasmer/Nextcloud/Zotero/Basu et al_2023_Equi-Tuning.pdf}
}
@@ -954,29 +1025,45 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Batatia, Ilyes and Batzner, Simon and Kovács, Dávid Péter and Musaelian, Albert and Simm, Gregor N. C. and Drautz, Ralf and Ortner, Christoph and Kozinsky, Boris and Csányi, Gábor},
date = {2022-05-13},
eprint = {2205.06643},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, stat},
- publisher = {arXiv},
doi = {10.48550/arXiv.2205.06643},
url = {http://arxiv.org/abs/2205.06643},
urldate = {2022-05-21},
abstract = {The rapid progress of machine learning interatomic potentials over the past couple of years produced a number of new architectures. Particularly notable among these are the Atomic Cluster Expansion (ACE), which unified many of the earlier ideas around atom density-based descriptors, and Neural Equivariant Interatomic Potentials (NequIP), a message passing neural network with equivariant features that showed state of the art accuracy. In this work, we construct a mathematical framework that unifies these models: ACE is generalised so that it can be recast as one layer of a multi-layer architecture. From another point of view, the linearised version of NequIP is understood as a particular sparsification of a much larger polynomial model. Our framework also provides a practical tool for systematically probing different choices in the unified design space. We demonstrate this by an ablation study of NequIP via a set of experiments looking at in- and out-of-domain accuracy and smooth extrapolation very far from the training data, and shed some light on which design choices are critical for achieving high accuracy. Finally, we present BOTNet (Body-Ordered-Tensor-Network), a much-simplified version of NequIP, which has an interpretable architecture and maintains accuracy on benchmark datasets.},
- keywords = {\_tablet,ACE,BOTNet,descriptors,equivariant,GNN,ML,MLP,MPNN,NequIP,NN,unified theory},
+ pubstate = {prepublished},
+ keywords = {ACE,BOTNet,descriptors,equivariant,GNN,MACE,ML,MLP,MPNN,NequIP,NN,unified theory},
file = {/Users/wasmer/Nextcloud/Zotero/Batatia et al_2022_The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentials.pdf;/Users/wasmer/Zotero/storage/2FLTPTA2/2205.html}
}
+@online{batatiaDesignSpaceEquivariant2022a,
+ title = {The {{Design Space}} of {{E}}(3)-{{Equivariant Atom-Centered Interatomic Potentials}}},
+ author = {Batatia, Ilyes and Batzner, Simon and Kovács, Dávid Péter and Musaelian, Albert and Simm, Gregor N. C. and Drautz, Ralf and Ortner, Christoph and Kozinsky, Boris and Csányi, Gábor},
+ date = {2022-11-24},
+ eprint = {2205.06643},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat, physics:physics, stat},
+ doi = {10.48550/arXiv.2205.06643},
+ url = {http://arxiv.org/abs/2205.06643},
+ urldate = {2024-06-17},
+ abstract = {The rapid progress of machine learning interatomic potentials over the past couple of years produced a number of new architectures. Particularly notable among these are the Atomic Cluster Expansion (ACE), which unified many of the earlier ideas around atom density-based descriptors, and Neural Equivariant Interatomic Potentials (NequIP), a message passing neural network with equivariant features that showed state of the art accuracy. In this work, we construct a mathematical framework that unifies these models: ACE is generalised so that it can be recast as one layer of a multi-layer architecture. From another point of view, the linearised version of NequIP is understood as a particular sparsification of a much larger polynomial model. Our framework also provides a practical tool for systematically probing different choices in the unified design space. We demonstrate this by an ablation study of NequIP via a set of experiments looking at in- and out-of-domain accuracy and smooth extrapolation very far from the training data, and shed some light on which design choices are critical for achieving high accuracy. Finally, we present BOTNet (Body-Ordered-Tensor-Network), a much-simplified version of NequIP, which has an interpretable architecture and maintains accuracy on benchmark datasets.},
+ pubstate = {prepublished},
+ keywords = {ACE,BOTNet,descriptors,equivariant,GNN,MACE,ML,MLP,MPNN,NequIP,NN,unified theory},
+ file = {/Users/wasmer/Nextcloud/Zotero/Batatia et al_2022_The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentials2.pdf;/Users/wasmer/Zotero/storage/KVDSWW6I/2205.html}
+}
+
@online{batatiaEquivariantMatrixFunction2023,
title = {Equivariant {{Matrix Function Neural Networks}}},
author = {Batatia, Ilyes and Schaaf, Lars L. and Chen, Huajie and Csányi, Gábor and Ortner, Christoph and Faber, Felix A.},
date = {2023-10-16},
eprint = {2310.10434},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, stat},
doi = {10.48550/arXiv.2310.10434},
url = {http://arxiv.org/abs/2310.10434},
urldate = {2023-11-05},
abstract = {Graph Neural Networks (GNNs), especially message-passing neural networks (MPNNs), have emerged as powerful architectures for learning on graphs in diverse applications. However, MPNNs face challenges when modeling non-local interactions in systems such as large conjugated molecules, metals, or amorphous materials. Although Spectral GNNs and traditional neural networks such as recurrent neural networks and transformers mitigate these challenges, they often lack extensivity, adaptability, generalizability, computational efficiency, or fail to capture detailed structural relationships or symmetries in the data. To address these concerns, we introduce Matrix Function Neural Networks (MFNs), a novel architecture that parameterizes non-local interactions through analytic matrix equivariant functions. Employing resolvent expansions offers a straightforward implementation and the potential for linear scaling with system size. The MFN architecture achieves state-of-the-art performance in standard graph benchmarks, such as the ZINC and TU datasets, and is able to capture intricate non-local interactions in quantum systems, paving the way to new state-of-the-art force fields.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ACE,AML,equivariant,linear scaling,long-range interaction,MACE,Metals and alloys,ML,ML-DFT,ML-ESM,MLP,MPNN,non-local interaction,prediction of Hamiltonian matrix,representation learning,spectral GNN,ZINC},
file = {/Users/wasmer/Nextcloud/Zotero/Batatia et al_2023_Equivariant Matrix Function Neural Networks.pdf;/Users/wasmer/Zotero/storage/2D2JTAAK/2310.html}
}
@@ -986,14 +1073,14 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Batatia, Ilyes and Benner, Philipp and Chiang, Yuan and Elena, Alin M. and Kovács, Dávid P. and Riebesell, Janosh and Advincula, Xavier R. and Asta, Mark and Baldwin, William J. and Bernstein, Noam and Bhowmik, Arghya and Blau, Samuel M. and Cărare, Vlad and Darby, James P. and De, Sandip and Della Pia, Flaviano and Deringer, Volker L. and Elijošius, Rokas and El-Machachi, Zakariya and Fako, Edvin and Ferrari, Andrea C. and Genreith-Schriever, Annalena and George, Janine and Goodall, Rhys E. A. and Grey, Clare P. and Han, Shuang and Handley, Will and Heenen, Hendrik H. and Hermansson, Kersti and Holm, Christian and Jaafar, Jad and Hofmann, Stephan and Jakob, Konstantin S. and Jung, Hyunwook and Kapil, Venkat and Kaplan, Aaron D. and Karimitari, Nima and Kroupa, Namu and Kullgren, Jolla and Kuner, Matthew C. and Kuryla, Domantas and Liepuoniute, Guoda and Margraf, Johannes T. and Magdău, Ioan-Bogdan and Michaelides, Angelos and Moore, J. Harry and Naik, Aakash A. and Niblett, Samuel P. and Norwood, Sam Walton and O'Neill, Niamh and Ortner, Christoph and Persson, Kristin A. and Reuter, Karsten and Rosen, Andrew S. and Schaaf, Lars L. and Schran, Christoph and Sivonxay, Eric and Stenczel, Tamás K. and Svahn, Viktor and Sutton, Christopher and family=Oord, given=Cas, prefix=van der, useprefix=true and Varga-Umbrich, Eszter and Vegge, Tejs and Vondrák, Martin and Wang, Yangshuai and Witt, William C. and Zills, Fabian and Csányi, Gábor},
date = {2023-12-29},
eprint = {2401.00096},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2401.00096},
url = {http://arxiv.org/abs/2401.00096},
urldate = {2024-01-20},
abstract = {Machine-learned force fields have transformed the atomistic modelling of materials by enabling simulations of ab initio quality on unprecedented time and length scales. However, they are currently limited by: (i) the significant computational and human effort that must go into development and validation of potentials for each particular system of interest; and (ii) a general lack of transferability from one chemical system to the next. Here, using the state-of-the-art MACE architecture we introduce a single general-purpose ML model, trained on a public database of 150k inorganic crystals, that is capable of running stable molecular dynamics on molecules and materials. We demonstrate the power of the MACE-MP-0 model -- and its qualitative and at times quantitative accuracy -- on a diverse set problems in the physical sciences, including the properties of solids, liquids, gases, and chemical reactions. The model can be applied out of the box and as a starting or "foundation model" for any atomistic system of interest and is thus a step towards democratising the revolution of ML force fields by lowering the barriers to entry.},
- pubstate = {preprint},
- keywords = {/unread,\_tablet,AML,benchmarking,foundation models,GNN,groundbreaking,MACE,MACE-MP-0,ML,MLP,MLP comparison,MPNN,original publication,PES,prediction of energy,pretrained models,SOTA,todo-tagging,universal potential,with-code},
+ pubstate = {prepublished},
+ keywords = {/unread,AML,benchmarking,foundation models,GNN,groundbreaking,MACE,MACE-MP-0,ML,MLP,MLP comparison,MPNN,original publication,PES,prediction of energy,pretrained models,SOTA,todo-tagging,universal potential,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Batatia et al_2023_A foundation model for atomistic materials chemistry.pdf;/Users/wasmer/Zotero/storage/M6QDTIVB/2401.html}
}
@@ -1002,14 +1089,14 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Batatia, Ilyes and Geiger, Mario and Munoz, Jose and Smidt, Tess and Silberman, Lior and Ortner, Christoph},
date = {2023-05-31},
eprint = {2306.00091},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {hep-th, stat},
doi = {10.48550/arXiv.2306.00091},
url = {http://arxiv.org/abs/2306.00091},
urldate = {2023-12-18},
abstract = {Reductive Lie Groups, such as the orthogonal groups, the Lorentz group, or the unitary groups, play essential roles across scientific fields as diverse as high energy physics, quantum mechanics, quantum chromodynamics, molecular dynamics, computer vision, and imaging. In this paper, we present a general Equivariant Neural Network architecture capable of respecting the symmetries of the finite-dimensional representations of any reductive Lie Group G. Our approach generalizes the successful ACE and MACE architectures for atomistic point clouds to any data equivariant to a reductive Lie group action. We also introduce the lie-nn software library, which provides all the necessary tools to develop and implement such general G-equivariant neural networks. It implements routines for the reduction of generic tensor products of representations into irreducible representations, making it easy to apply our architecture to a wide range of problems and groups. The generality and performance of our approach are demonstrated by applying it to the tasks of top quark decay tagging (Lorentz group) and shape recognition (orthogonal group).},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,AML,cluster expansion,E(3),equivariant,General ML,geometric deep learning,GNN,library,Lie groups,MACE,ML,MPNN,symmetry,with-code},
+ pubstate = {prepublished},
+ keywords = {ACE,AML,cluster expansion,E(3),equivariant,General ML,geometric deep learning,GNN,library,Lie groups,MACE,ML,MPNN,symmetry,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Batatia et al_2023_A General Framework for Equivariant Neural Networks on Reductive Lie Groups.pdf;/Users/wasmer/Zotero/storage/6X2PAYFQ/2306.html}
}
@@ -1019,17 +1106,34 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Batatia, Ilyes and Kovács, Dávid Péter and Simm, Gregor N. C. and Ortner, Christoph and Csányi, Gábor},
date = {2022-06-15},
eprint = {2206.07697},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, stat},
doi = {10.48550/arXiv.2206.07697},
- url = {http://arxiv.org/abs/2206.07697},
+ url = {http://arxiv.org/abs/2206.07697v1},
urldate = {2022-09-25},
abstract = {Creating fast and accurate force fields is a long-standing challenge in computational chemistry and materials science. Recently, several equivariant message passing neural networks (MPNNs) have been shown to outperform models built using other approaches in terms of accuracy. However, most MPNNs suffer from high computational cost and poor scalability. We propose that these limitations arise because MPNNs only pass two-body messages leading to a direct relationship between the number of layers and the expressivity of the network. In this work, we introduce MACE, a new equivariant MPNN model that uses higher body order messages. In particular, we show that using four-body messages reduces the required number of message passing iterations to just \textbackslash emph\{two\}, resulting in a fast and highly parallelizable model, reaching or exceeding state-of-the-art accuracy on the rMD17, 3BPA, and AcAc benchmark tasks. We also demonstrate that using higher order messages leads to an improved steepness of the learning curves.},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,chemical species scaling problem,descriptors,equivariant,library,MACE,ML,MLP,models,MPNN,Multi-ACE,NequIP,original publication,unified theory,with-code},
+ pubstate = {prepublished},
+ keywords = {ACE,chemical species scaling problem,descriptors,equivariant,library,MACE,ML,MLP,models,MPNN,Multi-ACE,NequIP,original publication,unified theory,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Batatia et al_2022_MACE.pdf;/Users/wasmer/Zotero/storage/LDAKZMRF/2206.html}
}
+@online{batatiaMACEHigherOrder2023,
+ title = {{{MACE}}: {{Higher Order Equivariant Message Passing Neural Networks}} for {{Fast}} and {{Accurate Force Fields}}},
+ shorttitle = {{{MACE}}},
+ author = {Batatia, Ilyes and Kovács, Dávid Péter and Simm, Gregor N. C. and Ortner, Christoph and Csányi, Gábor},
+ date = {2023-01-26},
+ eprint = {2206.07697},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat, physics:physics, stat},
+ doi = {10.48550/arXiv.2206.07697},
+ url = {http://arxiv.org/abs/2206.07697v2},
+ urldate = {2024-06-17},
+ abstract = {Creating fast and accurate force fields is a long-standing challenge in computational chemistry and materials science. Recently, several equivariant message passing neural networks (MPNNs) have been shown to outperform models built using other approaches in terms of accuracy. However, most MPNNs suffer from high computational cost and poor scalability. We propose that these limitations arise because MPNNs only pass two-body messages leading to a direct relationship between the number of layers and the expressivity of the network. In this work, we introduce MACE, a new equivariant MPNN model that uses higher body order messages. In particular, we show that using four-body messages reduces the required number of message passing iterations to just two, resulting in a fast and highly parallelizable model, reaching or exceeding state-of-the-art accuracy on the rMD17, 3BPA, and AcAc benchmark tasks. We also demonstrate that using higher order messages leads to an improved steepness of the learning curves.},
+ pubstate = {prepublished},
+ keywords = {ACE,chemical species scaling problem,descriptors,equivariant,library,MACE,ML,MLP,models,MPNN,Multi-ACE,NequIP,original publication,unified theory,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Batatia et al_2023_MACE.pdf;/Users/wasmer/Zotero/storage/LTHAFT4B/2206.html}
+}
+
@article{batraEmergingMaterialsIntelligence2020,
title = {Emerging Materials Intelligence Ecosystems Propelled by Machine Learning},
author = {Batra, Rohit and Song, Le and Ramprasad, Rampi},
@@ -1081,7 +1185,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
urldate = {2023-10-01},
abstract = {Topological insulators are a new class of materials that have attracted significant attention in contemporary condensed matter physics. They are different from regular insulators, and they display novel quantum properties that involve the idea of ‘topology’, an area of mathematics. Some of the fundamental concepts behind topological insulators, particularly in low-dimensional condensed matter systems such as poly-acetylene chains, can be understood using a simple one-dimensional toy model popularly known as the Su-Schrieffer-Heeger (SSH) model. This model can also be used as an introduction to the topological insulators of higher dimensions. Here, we give a concise description of the SSH model along with a brief review of the background physics and attempt to understand the ideas of topological invariants, edge states, and bulk-boundary correspondence using the model.},
langid = {english},
- keywords = {\_tablet,1D,Berry phase,condensed matter,educational,learning material,physics,Su-Schrieffer-Heeger model,TB,topological,topological insulator,tutorial},
+ keywords = {1D,Berry phase,condensed matter,educational,learning material,physics,Su-Schrieffer-Heeger model,TB,topological,topological insulator,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Batra_Sheet_2020_Physics with Coffee and Doughnuts.pdf}
}
@@ -1111,13 +1215,13 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Batzner, Simon and Musaelian, Albert and Sun, Lixin and Geiger, Mario and Mailoa, Jonathan P. and Kornbluth, Mordechai and Molinari, Nicola and Smidt, Tess E. and Kozinsky, Boris},
date = {2021-12-16},
eprint = {2101.03164},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2101.03164},
urldate = {2022-01-02},
abstract = {This work presents Neural Equivariant Interatomic Potentials (NequIP), an E(3)-equivariant neural network approach for learning interatomic potentials from ab-initio calculations for molecular dynamics simulations. While most contemporary symmetry-aware models use invariant convolutions and only act on scalars, NequIP employs E(3)-equivariant convolutions for interactions of geometric tensors, resulting in a more information-rich and faithful representation of atomic environments. The method achieves state-of-the-art accuracy on a challenging and diverse set of molecules and materials while exhibiting remarkable data efficiency. NequIP outperforms existing models with up to three orders of magnitude fewer training data, challenging the widely held belief that deep neural networks require massive training sets. The high data efficiency of the method allows for the construction of accurate potentials using high-order quantum chemical level of theory as reference and enables high-fidelity molecular dynamics simulations over long time scales.},
version = {3},
- keywords = {\_tablet,GNN,MD,ML,MLP,molecules,MPNN,NequIP,Neural networks,Physics - Computational Physics,solids},
+ keywords = {GNN,MD,ML,MLP,molecules,MPNN,NequIP,Neural networks,Physics - Computational Physics,solids},
file = {/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Batzner et al_2021_E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate.pdf;/Users/wasmer/Zotero/storage/85ATGPNR/s41467-022-29939-5.html;/Users/wasmer/Zotero/storage/V4Y8BWNW/2101.html}
}
@@ -1132,7 +1236,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
abstract = {This thesis is concerned with the quantum mechanical investigation of a novel class of magnetic phenomena in atomic- and nanoscale-sized systems deposited on surfaces or embedded in bulk materials that result from a competition between the exchange and the relativistic spin-orbit interactions. The thesis is motivated by the observation of novel spin-textures of one- and two-dimensional periodicity of nanoscale pitchlength exhibiting a unique winding sense observed in ultra-thin magnetic lms on nonmagnetic metallic substrates with a large spin-orbit interaction. The goal is to extend this eld to magnetic clusters and nano-structures of nite size in order to investigate in how far the size of the cluster and the atoms at the edge of the cluster or ribbon that are particular susceptible to relativistic eects change the balance betweendierent interactions and thus lead to new magnetic phenomena. As an example, the challenging problem of Fe nano-islands on Ir(111) is addressed in detail as for an Fe monolayer on Ir(111) a magnetic nanoskyrmion lattice was observed as magnetic structure.To achieve this goal a new rst-principles all-electron electronic structure code based on density functional theory was developed. The method of choice is the Korringa-Kohn-Rostoker (KKR) impurity Green function method, resorting on a multiple scattering approach. This method has been conceptually further advanced to combine the neglect of any shape approximation to the full potential, with the treatment ofnon-collinear magnetism, of the spin-orbit interaction, as well as of the structural relaxation together with the perfect embedding of a nite size magnetic cluster of atoms into a surface or a bulk environment. For this purpose the formalism makes use of an expansion of the Green function involving explicitly left- and right-hand side scattering solutions. Relativistic eects are treated via the scalar-relativistic approximation and a spin-orbit coupling term treated self-consistently. This required the development of a new algorithm to solve the relativistic quantum mechanical scattering problem for a single atom with a non-spherical potential formulated in terms of the Lippmann-Schwinger integral equation. Prior to the investigation of the Fe nano-islands, the magnetic structure of an Fe monolayer is studied using atomistic spin-dynamics on the basis of a classical model Hamiltonian, which uses realistic coupling parameters obtained from rst principles. It is shown that this method is capable to nd the experimentally determined magnetic structure. [...] Bauer, David Siegfried Georg},
isbn = {9783893369348},
langid = {english},
- keywords = {\_tablet,juKKR,KKR,KKRimp,PGI-1/IAS-1,thesis},
+ keywords = {juKKR,KKR,KKRimp,PGI-1/IAS-1,thesis},
file = {/Users/wasmer/Nextcloud/Zotero/Bauer_2014_Development of a relativistic full-potential first-principles multiple.pdf;/Users/wasmer/Zotero/storage/SYS2ZV93/151022.html}
}
@@ -1190,7 +1294,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
urldate = {2021-05-18},
abstract = {A lot of progress has been made in recent years in the development of atomistic potentials using machine learning (ML) techniques. In contrast to most conventional potentials, which are based on physical approximations and simplifications to derive an analytic functional relation between the atomic configuration and the potential-energy, ML potentials rely on simple but very flexible mathematical terms without a direct physical meaning. Instead, in case of ML potentials the topology of the potential-energy surface is “learned†by adjusting a number of parameters with the aim to reproduce a set of reference electronic structure data as accurately as possible. Due to this bias-free construction, they are applicable to a wide range of systems without changes in their functional form, and a very high accuracy close to the underlying first-principles data can be obtained. Neural network potentials (NNPs), which have first been proposed about two decades ago, are an important class of ML potentials. Although the first NNPs have been restricted to small molecules with only a few degrees of freedom, they are now applicable to high-dimensional systems containing thousands of atoms, which enables addressing a variety of problems in chemistry, physics, and materials science. In this tutorial review, the basic ideas of NNPs are presented with a special focus on developing NNPs for high-dimensional condensed systems. A recipe for the construction of these potentials is given and remaining limitations of the method are discussed. © 2015 Wiley Periodicals, Inc.},
langid = {english},
- keywords = {\_tablet,HDNNP,ML,models,molecular dynamics,neural network potentials,review,tutorial},
+ keywords = {HDNNP,ML,models,molecular dynamics,neural network potentials,review,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Behler_2015_Constructing high-dimensional neural network potentials.pdf;/Users/wasmer/Zotero/storage/DQEEE6BV/qua.html}
}
@@ -1220,7 +1324,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
number = {14},
doi = {10.1103/PhysRevLett.98.146401},
keywords = {BPNN,MD,ML,MLP,models,NN,original publication},
- file = {/Users/wasmer/Zotero/storage/RNTYUSXX/PhysRevLett.98.html}
+ file = {/Users/wasmer/Nextcloud/Zotero/Behler_2007_Generalized Neural-Network Representation of High-Dimensional Potential-Energy.pdf;/Users/wasmer/Zotero/storage/RNTYUSXX/PhysRevLett.98.html}
}
@article{behlerPerspectiveMachineLearning2016,
@@ -1251,7 +1355,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
volume = {123},
number = {4},
eprint = {2307.06879},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
pages = {042405},
issn = {0003-6951, 1077-3118},
@@ -1259,7 +1363,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
url = {http://arxiv.org/abs/2307.06879},
urldate = {2023-08-19},
abstract = {Technologies that function at room temperature often require magnets with a high Curie temperature, \$T\_\textbackslash mathrm\{C\}\$, and can be improved with better materials. Discovering magnetic materials with a substantial \$T\_\textbackslash mathrm\{C\}\$ is challenging because of the large number of candidates and the cost of fabricating and testing them. Using the two largest known data sets of experimental Curie temperatures, we develop machine-learning models to make rapid \$T\_\textbackslash mathrm\{C\}\$ predictions solely based on the chemical composition of a material. We train a random forest model and a \$k\$-NN one and predict on an initial dataset of over 2,500 materials and then validate the model on a new dataset containing over 3,000 entries. The accuracy is compared for multiple compounds' representations ("descriptors") and regression approaches. A random forest model provides the most accurate predictions and is not improved by dimensionality reduction or by using more complex descriptors based on atomic properties. A random forest model trained on a combination of both datasets shows that cobalt-rich and iron-rich materials have the highest Curie temperatures for all binary and ternary compounds. An analysis of the model reveals systematic error that causes the model to over-predict low-\$T\_\textbackslash mathrm\{C\}\$ materials and under-predict high-\$T\_\textbackslash mathrm\{C\}\$ materials. For exhaustive searches to find new high-\$T\_\textbackslash mathrm\{C\}\$ materials, analysis of the learning rate suggests either that much more data is needed or that more efficient descriptors are necessary.},
- keywords = {todo-tagging},
+ keywords = {\_tablet,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Belot et al_2023_Machine Learning Predictions of High-Curie-Temperature Materials.pdf;/Users/wasmer/Zotero/storage/R2ZPPEBA/2307.html}
}
@@ -1301,6 +1405,24 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
file = {/Users/wasmer/Nextcloud/Zotero/Bengio et al_2013_Representation Learning.pdf;/Users/wasmer/Zotero/storage/PEAGSIHD/6472238.html}
}
+@article{benmahmoudDataNextChallenge2024,
+ title = {Data as the next Challenge in Atomistic Machine Learning},
+ author = {Ben Mahmoud, Chiheb and Gardner, John L. A. and Deringer, Volker L.},
+ date = {2024-06-12},
+ journaltitle = {Nature Computational Science},
+ shortjournal = {Nat Comput Sci},
+ pages = {1--4},
+ publisher = {Nature Publishing Group},
+ issn = {2662-8457},
+ doi = {10.1038/s43588-024-00636-1},
+ url = {https://www.nature.com/articles/s43588-024-00636-1},
+ urldate = {2024-06-16},
+ abstract = {As machine learning models are becoming mainstream tools for molecular and materials research, there is an urgent need to improve the nature, quality, and accessibility of atomistic data. In turn, there are opportunities for a new generation of generally applicable datasets and distillable models.},
+ langid = {english},
+ keywords = {/unread,AI4Science,AML,commentary,data augmentation,database generation,foundation models,ML,ML-FPO,op-ed,roadmap,synthetic data},
+ file = {/Users/wasmer/Nextcloud/Zotero/Ben Mahmoud et al_2024_Data as the next challenge in atomistic machine learning.pdf}
+}
+
@article{benmahmoudLearningElectronicDensity2020,
title = {Learning the Electronic Density of States in Condensed Matter},
author = {Ben Mahmoud, Chiheb and Anelli, Andrea and Csányi, Gábor and Ceriotti, Michele},
@@ -1360,7 +1482,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Berner, Julius and Grohs, Philipp and Kutyniok, Gitta and Petersen, Philipp},
date = {2021-05-09},
eprint = {2105.04026},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
url = {http://arxiv.org/abs/2105.04026},
urldate = {2022-01-02},
@@ -1385,7 +1507,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
abstract = {Magnetic topological materials represent a class of compounds with properties that are strongly influenced by the topology of their electronic wavefunctions coupled with the magnetic spin configuration. Such materials can support chiral electronic channels of perfect conduction, and can be used for an array of applications, from information storage and control to dissipationless spin and charge transport. Here we review the theoretical and experimental progress achieved in the field of magnetic topological materials, beginning with the theoretical prediction of the quantum anomalous Hall effect without Landau levels, and leading to the recent discoveries of magnetic Weyl semimetals and antiferromagnetic topological insulators. We outline recent theoretical progress that has resulted in the tabulation of, for the first time, all magnetic symmetry group representations and topology. We describe several experiments realizing Chern insulators, Weyl and Dirac magnetic semimetals, and an array of axionic and higher-order topological phases of matter, and we survey future perspectives.},
issue = {7899},
langid = {english},
- keywords = {/unread,Topological matter},
+ keywords = {ARPES,Chern insulator,condensed matter,Dirac semimetal,experimental,Ferromagnetism,for introductions,Hall AHE,Hall effect,Hall QAHE,magnetism,physics,review,review-of-TIs,semimetal,STM,topological,topological insulator,Topological matter,Weyl semimetal},
file = {/Users/wasmer/Nextcloud/Zotero/Bernevig et al_2022_Progress and prospects in magnetic topological materials.pdf}
}
@@ -1404,7 +1526,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
urldate = {2023-06-14},
abstract = {In this article, we provide an overview of the basic concepts of novel topological materials. This new class of materials developed by combining the Weyl/Dirac fermionic electron states and magnetism, provide a materials-science platform to test predictions of the laws of topological physics. Owing to their dissipationless transport, these materials hold high promises for technological applications in quantum computing and spintronics devices.},
langid = {english},
- keywords = {/unread,\_tablet,ARPES,Berry phase,breaking of TRS,Fermi arc,Hall effect,Hall QHE,semimetal,TKNN,topological insulator,TRS},
+ keywords = {/unread,ARPES,Berry phase,breaking of TRS,Fermi arc,Hall effect,Hall QHE,semimetal,TKNN,topological insulator,TRS},
file = {/Users/wasmer/Nextcloud/Zotero/Bhardwaj_Chatterjee_2020_Topological Materials.pdf}
}
@@ -1423,7 +1545,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
url = {https://aip.scitation.org/doi/10.1063/5.0124363},
urldate = {2022-12-29},
abstract = {Machine learning frameworks based on correlations of interatomic positions begin with a discretized description of the density of other atoms in the neighborhood of each atom in the system. Symmetry considerations support the use of spherical harmonics to expand the angular dependence of this density, but there is, as of yet, no clear rationale to choose one radial basis over another. Here, we investigate the basis that results from the solution of the Laplacian eigenvalue problem within a sphere around the atom of interest. We show that this generates a basis of controllable smoothness within the sphere (in the same sense as plane waves provide a basis with controllable smoothness for a problem with periodic boundaries) and that a tensor product of Laplacian eigenstates also provides a smooth basis for expanding any higher-order correlation of the atomic density within the appropriate hypersphere. We consider several unsupervised metrics of the quality of a basis for a given dataset and show that the Laplacian eigenstate basis has a performance that is much better than some widely used basis sets and competitive with data-driven bases that numerically optimize each metric. Finally, we investigate the role of the basis in building models of the potential energy. In these tests, we find that a combination of the Laplacian eigenstate basis and target-oriented heuristics leads to equal or improved regression performance when compared to both heuristic and data-driven bases in the literature. We conclude that the smoothness of the basis functions is a key aspect of successful atomic density representations.},
- keywords = {\_tablet,ACDC,ACE,density correlation,descriptor comparison,descriptors,equivariant,general body-order,GPR,Jacobian condition number,Laplacian eigenstate basis,MACE,NN,prediction of potential energy,radial basis,regression,residual variance,smooth basis,Supervised learning,unsupervised learning},
+ keywords = {ACDC,ACE,density correlation,descriptor comparison,descriptors,equivariant,general body-order,GPR,Jacobian condition number,Laplacian eigenstate basis,MACE,NN,prediction of potential energy,radial basis,regression,residual variance,smooth basis,Supervised learning,unsupervised learning},
file = {/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Bigi et al_2022_A smooth basis for atomistic machine learning.pdf}
}
@@ -1433,14 +1555,14 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Bigi, Filippo and Pozdnyakov, Sergey N. and Ceriotti, Michele},
date = {2023-03-07},
eprint = {2303.04124},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2303.04124},
url = {http://arxiv.org/abs/2303.04124},
urldate = {2023-04-13},
abstract = {Machine-learning models based on a point-cloud representation of a physical object are ubiquitous in scientific applications and particularly well-suited to the atomic-scale description of molecules and materials. Among the many different approaches that have been pursued, the description of local atomic environments in terms of their neighbor densities has been used widely and very succesfully. We propose a novel density-based method which involves computing ``Wigner kernels''. These are fully equivariant and body-ordered kernels that can be computed iteratively with a cost that is independent of the radial-chemical basis and grows only linearly with the maximum body-order considered. This is in marked contrast to feature-space models, which comprise an exponentially-growing number of terms with increasing order of correlations. We present several examples of the accuracy of models based on Wigner kernels in chemical applications, for both scalar and tensorial targets, reaching state-of-the-art accuracy on the popular QM9 benchmark dataset, and we discuss the broader relevance of these ideas to equivariant geometric machine-learning.},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,Allegro,AML,benchmarking,body-order,descriptor comparison,descriptors,DimeNet++,equivariant,GPR,KRR,lambda-SOAP,ML,MLP,model comparison,molecules,MPNN,NICE,PAiNN,QM9,representation learning,SA-GPR,SOAP,SphereNet,tensorial target,Wigner kernel},
+ pubstate = {prepublished},
+ keywords = {ACE,Allegro,AML,benchmarking,body-order,descriptor comparison,descriptors,DimeNet++,equivariant,GPR,KRR,lambda-SOAP,ML,MLP,model comparison,molecules,MPNN,NICE,PAiNN,QM9,representation learning,SA-GPR,SOAP,SphereNet,tensorial target,Wigner kernel},
file = {/Users/wasmer/Nextcloud/Zotero/Bigi et al_2023_Wigner kernels.pdf;/Users/wasmer/Zotero/storage/LERSCPN4/2303.html}
}
@@ -1617,7 +1739,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
abstract = {Blügel, S.},
isbn = {9783893364305},
langid = {english},
- keywords = {\_tablet,AIMD,bluegel,DFT,FLEUR,IFF,IFF spring school,introduction,learning material,magnetism,PGI-1/IAS-1,rec-by-bluegel},
+ keywords = {AIMD,bluegel,DFT,FLEUR,IFF,IFF spring school,introduction,learning material,magnetism,PGI-1/IAS-1,rec-by-bluegel},
file = {/Users/wasmer/Nextcloud/Zotero/Blügel_2006_Density Functional Theory in Practice.pdf;/Users/wasmer/Zotero/storage/ZL4WZAY7/51316.html}
}
@@ -1646,7 +1768,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
abstract = {Condensed matter physics is currently undergoing a revolution through the introduction of concepts arising from topology that are used to characterize physical states, fields and properties from a completely different perspective. With the introduction of topology, the perspective is changed from describing complex systems in terms of local order parameters to a characterization by global quantities, which are measured nonlocally and which endow the systems with a global stability to perturbations. Prominent examples are topological insulators, skyrmions and Majorana fermions. Since topology translates into quantization, and topological order to entanglement, this ongoing revolution has impact on fields like mathematics, materials science, nanoelectronics and quantum information resulting in new device concepts enabling computations without dissipation of energy or enabling the possibility of realizing platforms for topological quantum computation, and ultimately reaching out into applications. Thus, these new exciting scientific developments and their applications are closely related to the grand challenges in information and communication technology and energy saving. Topology is the branch of mathematics that deals with properties of spaces that are invariant under smooth deformations. It provides newly appreciated mathematical tools in condensed matter physics that are currently revolutionizing the field of quantum matter and materials. Topology dictates that if two different Hamiltonians can be smoothly deformed into each other they give rise to many common physical properties and their states are homotopy invariant. Thus, topological invariance, which is often protected by discrete symmetries, provides some robustness that translates into the quantization of properties; such a robust quantization motivates the search and discovery of new topological matter. So far, the mainstream of modern topological condensed matter physics relies on two profoundly different scenarios: the emergence of the complex topology either in real space, as manifested e.g. in non-trivial magnetic structures or in momentum space, finding its realization in such materials as topological and Chern insulators. The latter renowned class of solids attracted considerable attention in recent years owing to its fascinating properties of spin-momentum locking, emergence of topologically protected surface/edge states governed by Dirac physics, as well as the quantization of Hall conductance and the discovery of the quantum spin Hall effect. Historically, the discovery of topological insulators gave rise to the discovery of a whole plethora of topologically non-trivial materials such asWeyl semimetals or topological superconductors, relevant in the context of the realization of Majorana fermions and topological quantum computation. [...]},
eventtitle = {Lecture {{Notes}} of the 48th {{IFF Spring School}} 2017},
isbn = {978-3-95806-202-3},
- keywords = {\_tablet,Berry phase,Chern insulator,Chern number,DFT,FZJ,Hall AHE,Hall effect,Hall QAHE,Hall QHE,Hall QSHE,Heisenberg model,IFF,IFF spring school,learning material,magnetic interactions,magnetic materials,magnetic topological materials,Majorana,MZM,PGI,PGI-1/IAS-1,PGI-9,quantum computing,review,skyrmions,spin-dependent,topological,topological insulator,tutorial},
+ keywords = {Berry phase,Chern insulator,Chern number,DFT,FZJ,Hall AHE,Hall effect,Hall QAHE,Hall QHE,Hall QSHE,Heisenberg model,IFF,IFF spring school,learning material,magnetic interactions,magnetic materials,magnetic topological materials,Majorana,MZM,PGI,PGI-1/IAS-1,PGI-9,quantum computing,review,skyrmions,spin-dependent,topological,topological insulator,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Blügel et al_2017_Topological Matter - Topological Insulators, Skyrmions and Majoranas.pdf}
}
@@ -1655,13 +1777,13 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Blum-Smith, Ben and Villar, Soledad},
date = {2023-03-25},
eprint = {2209.14991},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.2209.14991},
url = {http://arxiv.org/abs/2209.14991},
urldate = {2023-06-30},
abstract = {Inspired by constraints from physical law, equivariant machine learning restricts the learning to a hypothesis class where all the functions are equivariant with respect to some group action. Irreducible representations or invariant theory are typically used to parameterize the space of such functions. In this article, we introduce the topic and explain a couple of methods to explicitly parameterize equivariant functions that are being used in machine learning applications. In particular, we explicate a general procedure, attributed to Malgrange, to express all polynomial maps between linear spaces that are equivariant under the action of a group \$G\$, given a characterization of the invariant polynomials on a bigger space. The method also parametrizes smooth equivariant maps in the case that \$G\$ is a compact Lie group.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Clebsch-Gordan,equivariant,General ML,GNN,group theory,invariance,invariant theory,ML,ML theory,symmetry,tensor product},
file = {/Users/wasmer/Nextcloud/Zotero/Blum-Smith_Villar_2023_Machine learning and invariant theory.pdf;/Users/wasmer/Zotero/storage/K6DZQ29J/2209.html}
}
@@ -1680,6 +1802,22 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
file = {/home/johannes/Books/data_science/general_theory/Blum_FoundationsOfDataScience_1e-2020.pdf}
}
+@article{blumRoadmapMethodsSoftware2024,
+ title = {Roadmap on Methods and Software for Electronic Structure Based Simulations in Chemistry and Materials},
+ author = {Blum, Volker and Asahi, Ryoji and Autschbach, Jochen and Bannwarth, Christoph and Bihlmayer, Gustav and Blügel, Stefan and Burns, Lori A. and Crawford, T. Daniel and Dawson, William and family=Jong, given=Wibe Albert, prefix=de, useprefix=true and Draxl, Claudia and Filippi, Claudia and Genovese, Luigi and Giannozzi, Paolo and Govind, Niranjan and Hammes-Schiffer, Sharon and Hammond, Jeff R. and Hourahine, Benjamin and Jain, Anubhav and Kanai, Yosuke and Kent, Paul R C and Larsen, Ask Hjorth and Lehtola, Susi and Li, Xiaosong and Lindh, Roland and Maeda, Satoshi and Makri, Nancy and Moussa, Jonathan and Nakajima, Takahito and Nash, Jessica A. and Oliveira, Micael J. T. and Patel, Pansy D. and Pizzi, Giovanni and Pourtois, Geoffrey and Pritchard, Benjamin P. and Rabani, Eran and Reiher, Markus and Reining, Lucia and Ren, Xinguo and Rossi, Mariana and Schlegel, H. Bernhard and Seriani, Nicola and Slipchenko, Lyudmila V. and Thom, Alexander and Valeev, Edward F. and Van Troeye, Benoit and Visscher, Lucas and Vlcek, Vojtech and Werner, Hans-Joachim and Williams-Young, David B. and Windus, Theresa},
+ date = {2024},
+ journaltitle = {Electronic Structure},
+ shortjournal = {Electron. Struct.},
+ issn = {2516-1075},
+ doi = {10.1088/2516-1075/ad48ec},
+ url = {http://iopscience.iop.org/article/10.1088/2516-1075/ad48ec},
+ urldate = {2024-06-07},
+ abstract = {Contents 1. Introduction- Methods and software for electronic structure based simulations of chemistry and materials 2. Density Functional Theory: Formalism and Current Directions 3. Density functional methods - implementation, challenges, successes 4. Green’s function based many-body perturbation theory 5. Wave-function theory approaches – explicit approaches to electron correlation 6. Quantum Monte Carlo and stochastic electronic structure methods 7. Heavy element relativity, spin-orbit physics, and magnetism 8. Semiempirical methods 9. Simulating Nuclear Dynamics with Quantum Effects 10. Real-Time Propagation in Electronic Structure Theory 11. Spectroscopy 12. Tools for exploring potential energy surfaces 13. Managing complex computational workflows 14. Current and Future Computer Architectures 15. Electronic structure software engineering 16. Education and Training in Electronic Structure Theory: Navigating an Evolving Landscape 17. Electronic structure theory facing industry and realistic modeling of experiments 18. List of Acronyms},
+ langid = {english},
+ keywords = {DFT,FZJ,PGI,PGI-1/IAS-1,review-of-DFT,roadmap},
+ file = {/Users/wasmer/Nextcloud/Zotero/Blum et al_2024_Roadmap on methods and software for electronic structure based simulations in2.pdf}
+}
+
@book{blundellMagnetismCondensedMatter2001,
title = {Magnetism in Condensed Matter},
author = {Blundell, Stephen},
@@ -1691,27 +1829,11 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
abstract = {An understanding of the quantum mechanical nature of magnetism has led to the development of new magnetic materials which are used as permanent magnets, sensors, and information storage. Behind these practical applications lie a range of fundamental ideas, including symmetry breaking, order parameters, excitations, frustration, and reduced dimensionality. This superb new textbook presents a logical account of these ideas, staring from basic concepts in electromagnetsim and quantum mechanics. It outlines the origin of magnetic moments in atoms and how these moments can be affected by their local environment inside a crystal. The different types of interactions which can be present between magnetic moments are described. The final chapters of the book are devoted to the magnetic properties of metals, and to the complex behaviour which can occur when competing magnetic interactions are present and/or the system has a reduced dimensionality. Throughout the text, the theoretical principles are applied to real systems. There is substantial discussion of experimental techniques and current research topics.; The book is copiously illustrated and contains detailed appendices which cover the fundamental principles.},
isbn = {9780585483603 9781280375132 9780191586644 9786610375134 9780198505921},
langid = {english},
- keywords = {\_tablet,condensed matter,magnetism,textbook,undergraduate},
+ keywords = {condensed matter,magnetism,textbook,undergraduate},
annotation = {OCLC: 53956469},
file = {/Users/wasmer/Nextcloud/Zotero/Blundell_2001_Magnetism in condensed matter.pdf}
}
-@online{bochkarevAtomicClusterExpansion2023,
- title = {Atomic {{Cluster Expansion}} for Semilocal Interactions beyond Equivariant Message Passing},
- author = {Bochkarev, Anton and Lysogorskiy, Yury and Drautz, Ralf},
- date = {2023-11-27},
- eprint = {2311.16326},
- eprinttype = {arxiv},
- eprintclass = {cond-mat},
- doi = {10.48550/arXiv.2311.16326},
- url = {http://arxiv.org/abs/2311.16326},
- urldate = {2023-12-18},
- abstract = {We extend the basis functions of the Atomic Cluster Expansion to graphs. This naturally leads to a representation that enables us to describe semilocal interactions in physiscally and chemically transparent form. Simplifications of the graph Atomic Cluster Expansion recover the currently most accurate message-passing representations of atomic interactions. We demonstrate the accuracy and efficiency of our expansion for a number of small molecules, clusters and a general-purpose model for carbon.},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,AML,carbon,clusters,descriptors,equivariant,GNN,graph ACE,ML,MLP,MPNN,semilocal interactions,smal organic molecules,symmetry},
- file = {/Users/wasmer/Nextcloud/Zotero/Bochkarev et al_2023_Atomic Cluster Expansion for semilocal interactions beyond equivariant message.pdf;/Users/wasmer/Zotero/storage/UMGXF4LV/2311.html}
-}
-
@article{bochkarevEfficientParametrizationAtomic2022,
title = {Efficient Parametrization of the Atomic Cluster Expansion},
author = {Bochkarev, Anton and Lysogorskiy, Yury and Menon, Sarath and Qamar, Minaam and Mrovec, Matous and Drautz, Ralf},
@@ -1726,22 +1848,38 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
url = {https://link.aps.org/doi/10.1103/PhysRevMaterials.6.013804},
urldate = {2023-12-08},
abstract = {The atomic cluster expansion (ACE) provides a general, local, and complete representation of atomic energies. Here we present an efficient framework for parametrization of ACE models for elements, alloys, and molecules. To this end, we first introduce general requirements for a physically meaningful description of the atomic interaction, in addition to the usual equivariance requirements. We then demonstrate that ACE can be converged systematically with respect to two fundamental characteristics—the number and complexity of basis functions and the choice of nonlinear representation. The construction of ACE parametrizations is illustrated for several representative examples with different bond chemistries, including metallic copper, covalent carbon, and several multicomponent molecular and alloy systems. We discuss the Pareto front of optimal force to energy matching contributions in the loss function, the influence of regularization, the importance of consistent and reliable reference data, and the necessity of unbiased validation. Our ACE parametrization strategy is implemented in the freely available software package pacemaker that enables largely automated and GPU accelerated training. The resulting ACE models are shown to be superior or comparable to the best currently available ML potentials and can be readily used in large-scale atomistic simulations.},
- keywords = {\_tablet,ACDC,ACE,AML,descriptors,library,ML,pacemaker,with-code},
+ keywords = {ACDC,ACE,AML,descriptors,library,ML,pacemaker,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Bochkarev et al_2022_Efficient parametrization of the atomic cluster expansion.pdf;/Users/wasmer/Zotero/storage/4TU3DH2J/PhysRevMaterials.6.html}
}
+@online{bochkarevGraphAtomicCluster2024,
+ title = {Graph {{Atomic Cluster Expansion}} for Semilocal Interactions beyond Equivariant Message Passing},
+ author = {Bochkarev, Anton and Lysogorskiy, Yury and Drautz, Ralf},
+ date = {2024-01-20},
+ eprint = {2311.16326},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat},
+ doi = {10.48550/arXiv.2311.16326},
+ url = {http://arxiv.org/abs/2311.16326},
+ urldate = {2024-05-23},
+ abstract = {The Atomic Cluster Expansion provides local, complete basis functions that enable efficient parametrization of many-atom interactions. We extend the Atomic Cluster Expansion to incorporate graph basis functions. This naturally leads to representations that enable the efficient description of semilocal interactions in physically and chemically transparent form. Simplification of the graph expansion by tensor decomposition results in an iterative procedure that comprises current message-passing machine learning interatomic potentials. We demonstrate the accuracy and efficiency of the graph Atomic Cluster Expansion for a number of small molecules, clusters and a general-purpose model for carbon. We further show that the graph Atomic Cluster Expansion scales linearly with number of neighbors and layer depth of the graph basis functions.},
+ pubstate = {prepublished},
+ keywords = {ACE,AML,carbon,clusters,descriptors,equivariant,GNN,graph ACE,ML,MLP,MPNN,semilocal interactions,smal organic molecules,symmetry},
+ file = {/Users/wasmer/Nextcloud/Zotero/Bochkarev et al_2024_Graph Atomic Cluster Expansion for semilocal interactions beyond equivariant.pdf;/Users/wasmer/Zotero/storage/VRLXR3D3/2311.html}
+}
+
@online{bochkarevMultilayerAtomicCluster2022a,
title = {Multilayer Atomic Cluster Expansion for Semi-Local Interactions},
author = {Bochkarev, Anton and Lysogorskiy, Yury and Ortner, Christoph and Csányi, Gábor and Drautz, Ralf},
date = {2022-05-17},
eprint = {2205.08177},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2205.08177},
url = {http://arxiv.org/abs/2205.08177},
urldate = {2022-09-29},
abstract = {Traditionally, interatomic potentials assume local bond formation supplemented by long-range electrostatic interactions when necessary. This ignores intermediate range multi-atom interactions that arise from the relaxation of the electronic structure. Here, we present the multilayer atomic cluster expansion (ml-ACE) that includes collective, semi-local multi-atom interactions naturally within its remit. We demonstrate that ml-ACE significantly improves fit accuracy compared to a local expansion on selected examples and provide physical intuition to understand this improvement.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ACE,descriptors,ML,ml-ACE,MLP,multilayer-ACE,semilocal interactions},
file = {/Users/wasmer/Nextcloud/Zotero/Bochkarev et al_2022_Multilayer atomic cluster expansion for semi-local interactions2.pdf;/Users/wasmer/Zotero/storage/ZVU3IARD/2205.html}
}
@@ -1785,14 +1923,14 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
author = {Bogojeski, Mihail and Brockherde, Felix and Vogt-Maranto, Leslie and Li, Li and Tuckerman, Mark E. and Burke, Kieron and Müller, Klaus-Robert},
date = {2018-11-15},
eprint = {1811.06255},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.1811.06255},
url = {http://arxiv.org/abs/1811.06255},
urldate = {2022-07-08},
abstract = {The Kohn-Sham scheme of density functional theory is one of the most widely used methods to solve electronic structure problems for a vast variety of atomistic systems across different scientific fields. While the method is fast relative to other first principles methods and widely successful, the computational time needed is still not negligible, making it difficult to perform calculations for very large systems or over long time-scales. In this submission, we revisit a machine learning model capable of learning the electron density and the corresponding energy functional based on a set of training examples. It allows us to bypass solving the Kohn-Sham equations, providing a significant decrease in computation time. We specifically focus on the machine learning formulation of the Hohenberg-Kohn map and its decomposability. We give results and discuss challenges, limits and future directions.},
- pubstate = {preprint},
- keywords = {\_tablet,DFT,HK map,ML,ML-DFT,ML-ESM,ML-HK map,molecules,prediction of electron density},
+ pubstate = {prepublished},
+ keywords = {DFT,HK map,ML,ML-DFT,ML-ESM,ML-HK map,molecules,prediction of electron density},
file = {/Users/wasmer/Nextcloud/Zotero/Bogojeski et al_2018_Efficient prediction of 3D electron densities using machine learning.pdf;/Users/wasmer/Zotero/storage/MCBT39D4/1811.html}
}
@@ -1813,7 +1951,7 @@ Subject\_term\_id: electronic-properties-and-materials;phase-transitions-and-cri
abstract = {Kohn-Sham density functional theory (DFT) is a standard tool in most branches of chemistry, but accuracies for many molecules are limited to 2-3 kcal~⋅~mol−1 with presently-available functionals. Ab initio methods, such as coupled-cluster, routinely produce much higher accuracy, but computational costs limit their application to small molecules. In this paper, we leverage machine learning to calculate coupled-cluster energies from DFT densities, reaching quantum chemical accuracy (errors below 1 kcal~⋅~mol−1) on test data. Moreover, density-based Δ-learning (learning only the correction to a standard DFT calculation, termed Δ-DFT~) significantly reduces the amount of training data required, particularly when molecular symmetries are included. The robustness of Δ-DFT~ is highlighted by correcting “on the fly†DFT-based molecular dynamics (MD) simulations of resorcinol (C6H4(OH)2) to obtain MD trajectories with coupled-cluster accuracy. We conclude, therefore, that Δ-DFT~ facilitates running gas-phase MD simulations with quantum chemical accuracy, even for strained geometries and conformer changes where standard DFT fails.},
issue = {1},
langid = {english},
- keywords = {\_tablet,2-step model,AML,CC,CCSD(T),coupled cluster,Delta,delta learning,DFT,HK map,KKR,ML,ML-DFA,ML-DFT,ML-ESM,ML-HK map,molecules,multi-step model,prediction of electron density,with-code,Δ-machine learning},
+ keywords = {2-step model,AML,CC,CCSD(T),coupled cluster,Delta,delta learning,DFT,HK map,KKR,ML,ML-DFA,ML-DFT,ML-ESM,ML-HK map,molecules,multi-step model,prediction of electron density,with-code,Δ-machine learning},
annotation = {Bandiera\_abtest: a\\
Cc\_license\_type: cc\_by\\
Cg\_type: Nature Research Journals\\
@@ -1823,18 +1961,36 @@ Subject\_term\_id: computational-chemistry;computational-science},
file = {/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Bogojeski et al_2020_Quantum chemical accuracy from density functional approximations via machine.pdf}
}
+@article{bohacekArtPracticeStructurebased1996,
+ title = {The Art and Practice of Structure-Based Drug Design: {{A}} Molecular Modeling Perspective},
+ shorttitle = {The Art and Practice of Structure-Based Drug Design},
+ author = {Bohacek, Regine S. and McMartin, Colin and Guida, Wayne C.},
+ date = {1996},
+ journaltitle = {Medicinal Research Reviews},
+ volume = {16},
+ number = {1},
+ pages = {3--50},
+ issn = {1098-1128},
+ doi = {10.1002/(SICI)1098-1128(199601)16:1<3::AID-MED1>3.0.CO;2-6},
+ url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/%28SICI%291098-1128%28199601%2916%3A1%3C3%3A%3AAID-MED1%3E3.0.CO%3B2-6},
+ urldate = {2024-08-02},
+ langid = {english},
+ keywords = {chemical space,chemical space size estimate,chemistry,for introductions,materials informatics,organic chemistry,original publication},
+ file = {/Users/wasmer/Zotero/storage/5QB32HF4/(SICI)1098-1128(199601)1613AID-MED13.0.html}
+}
+
@online{bommasaniOpportunitiesRisksFoundation2022,
title = {On the {{Opportunities}} and {{Risks}} of {{Foundation Models}}},
author = {Bommasani, Rishi and Hudson, Drew A. and Adeli, Ehsan and Altman, Russ and Arora, Simran and family=Arx, given=Sydney, prefix=von, useprefix=true and Bernstein, Michael S. and Bohg, Jeannette and Bosselut, Antoine and Brunskill, Emma and Brynjolfsson, Erik and Buch, Shyamal and Card, Dallas and Castellon, Rodrigo and Chatterji, Niladri and Chen, Annie and Creel, Kathleen and Davis, Jared Quincy and Demszky, Dora and Donahue, Chris and Doumbouya, Moussa and Durmus, Esin and Ermon, Stefano and Etchemendy, John and Ethayarajh, Kawin and Fei-Fei, Li and Finn, Chelsea and Gale, Trevor and Gillespie, Lauren and Goel, Karan and Goodman, Noah and Grossman, Shelby and Guha, Neel and Hashimoto, Tatsunori and Henderson, Peter and Hewitt, John and Ho, Daniel E. and Hong, Jenny and Hsu, Kyle and Huang, Jing and Icard, Thomas and Jain, Saahil and Jurafsky, Dan and Kalluri, Pratyusha and Karamcheti, Siddharth and Keeling, Geoff and Khani, Fereshte and Khattab, Omar and Koh, Pang Wei and Krass, Mark and Krishna, Ranjay and Kuditipudi, Rohith and Kumar, Ananya and Ladhak, Faisal and Lee, Mina and Lee, Tony and Leskovec, Jure and Levent, Isabelle and Li, Xiang Lisa and Li, Xuechen and Ma, Tengyu and Malik, Ali and Manning, Christopher D. and Mirchandani, Suvir and Mitchell, Eric and Munyikwa, Zanele and Nair, Suraj and Narayan, Avanika and Narayanan, Deepak and Newman, Ben and Nie, Allen and Niebles, Juan Carlos and Nilforoshan, Hamed and Nyarko, Julian and Ogut, Giray and Orr, Laurel and Papadimitriou, Isabel and Park, Joon Sung and Piech, Chris and Portelance, Eva and Potts, Christopher and Raghunathan, Aditi and Reich, Rob and Ren, Hongyu and Rong, Frieda and Roohani, Yusuf and Ruiz, Camilo and Ryan, Jack and Ré, Christopher and Sadigh, Dorsa and Sagawa, Shiori and Santhanam, Keshav and Shih, Andy and Srinivasan, Krishnan and Tamkin, Alex and Taori, Rohan and Thomas, Armin W. and Tramèr, Florian and Wang, Rose E. and Wang, William and Wu, Bohan and Wu, Jiajun and Wu, Yuhuai and Xie, Sang Michael and Yasunaga, Michihiro and You, Jiaxuan and Zaharia, Matei and Zhang, Michael and Zhang, Tianyi and Zhang, Xikun and Zhang, Yuhui and Zheng, Lucia and Zhou, Kaitlyn and Liang, Percy},
date = {2022-07-12},
eprint = {2108.07258},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2108.07258},
url = {http://arxiv.org/abs/2108.07258},
urldate = {2023-04-14},
abstract = {AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles(e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities,and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {few-shot learning,foundation models,General ML,ML,transfer learning,transformer,zero-shot learning},
file = {/Users/wasmer/Nextcloud/Zotero/Bommasani et al_2022_On the Opportunities and Risks of Foundation Models.pdf;/Users/wasmer/Zotero/storage/72DPHWW4/2108.html}
}
@@ -1845,13 +2001,13 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Bondesan, Roberto and Welling, Max},
date = {2021-03-08},
eprint = {2103.04913},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {quant-ph},
doi = {10.48550/arXiv.2103.04913},
url = {http://arxiv.org/abs/2103.04913},
urldate = {2023-08-22},
abstract = {In this work we develop a quantum field theory formalism for deep learning, where input signals are encoded in Gaussian states, a generalization of Gaussian processes which encode the agent's uncertainty about the input signal. We show how to represent linear and non-linear layers as unitary quantum gates, and interpret the fundamental excitations of the quantum model as particles, dubbed ``Hintons''. On top of opening a new perspective and techniques for studying neural networks, the quantum formulation is well suited for optical quantum computing, and provides quantum deformations of neural networks that can be run efficiently on those devices. Finally, we discuss a semi-classical limit of the quantum deformed models which is amenable to classical simulation.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Bondesan_Welling_2021_The Hintons in your Neural Network.pdf;/Users/wasmer/Zotero/storage/E2RNIICV/2103.html}
}
@@ -1894,13 +2050,13 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Bosoni, Emanuele and Beal, Louis and Bercx, Marnik and Blaha, Peter and Blügel, Stefan and Bröder, Jens and Callsen, Martin and Cottenier, Stefaan and Degomme, Augustin and Dikan, Vladimir and Eimre, Kristjan and Flage-Larsen, Espen and Fornari, Marco and Garcia, Alberto and Genovese, Luigi and Giantomassi, Matteo and Huber, Sebastiaan P. and Janssen, Henning and Kastlunger, Georg and Krack, Matthias and Kresse, Georg and Kühne, Thomas D. and Lejaeghere, Kurt and Madsen, Georg K. H. and Marsman, Martijn and Marzari, Nicola and Michalicek, Gregor and Mirhosseini, Hossein and Müller, Tiziano M. A. and Petretto, Guido and Pickard, Chris J. and Poncé, Samuel and Rignanese, Gian-Marco and Rubel, Oleg and Ruh, Thomas and Sluydts, Michael and Vanpoucke, Danny E. P. and Vijay, Sudarshan and Wolloch, Michael and Wortmann, Daniel and Yakutovich, Aliaksandr V. and Yu, Jusong and Zadoks, Austin and Zhu, Bonan and Pizzi, Giovanni},
date = {2023-05-26},
eprint = {2305.17274},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2305.17274},
url = {http://arxiv.org/abs/2305.17274},
urldate = {2023-06-29},
abstract = {In the past decades many density-functional theory methods and codes adopting periodic boundary conditions have been developed and are now extensively used in condensed matter physics and materials science research. Only in 2016, however, their precision (i.e., to which extent properties computed with different codes agree among each other) was systematically assessed on elemental crystals: a first crucial step to evaluate the reliability of such computations. We discuss here general recommendations for verification studies aiming at further testing precision and transferability of density-functional-theory computational approaches and codes. We illustrate such recommendations using a greatly expanded protocol covering the whole periodic table from Z=1 to 96 and characterizing 10 prototypical cubic compounds for each element: 4 unaries and 6 oxides, spanning a wide range of coordination numbers and oxidation states. The primary outcome is a reference dataset of 960 equations of state cross-checked between two all-electron codes, then used to verify and improve nine pseudopotential-based approaches. Such effort is facilitated by deploying AiiDA common workflows that perform automatic input parameter selection, provide identical input/output interfaces across codes, and ensure full reproducibility. Finally, we discuss the extent to which the current results for total energies can be reused for different goals (e.g., obtaining formation energies).},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ABINIT,AiiDA,AiiDA-FLEUR,benchmarking,best practices,BigDFT,CASTEP,CP2K,DFT,DFT codes comparison,EOS,error estimate,FAIR,FLAPW,FLEUR,GPAW,guidelines,LAPW,library,Materials Cloud,oxides,PAW,pseudopotential,Quantum ESPRESSO,reference dataset,reproducibility,scientific workflows,SIESTA,SIRIUS,VASP,verification,WIEN2k,with-code,workflows},
file = {/Users/wasmer/Nextcloud/Zotero/Bosoni et al_2023_How to verify the precision of density-functional-theory implementations via.pdf;/Users/wasmer/Zotero/storage/GIP5Z2MT/2305.html}
}
@@ -1915,7 +2071,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
volume = {3},
number = {5},
doi = {10.1103/PhysRevMaterials.3.054201},
- keywords = {\_tablet,defects,Funsilab,impurity embedding,PGI-1/IAS-1,topological insulator},
+ keywords = {defects,Funsilab,impurity embedding,PGI-1/IAS-1,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Bouaziz_2019_Spin dynamics of 3d and 4d impurities embedded in prototypical topological.pdf;/Users/wasmer/Zotero/storage/CW3GMSS2/PhysRevMaterials.3.html}
}
@@ -1931,7 +2087,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
abstract = {This thesis provides a theoretical description of magnetic nanostructures in inversion-asymmetric environments with strong spin-orbit interaction (SOI). The theoretical concepts introduced here can be applied in the field of spin-orbitronics, which consists ofexploiting the SOI to manipulate the electron spin without external magnetic fields. The investigated systems display a plethora of interesting phenomena ranging from chiral magnetic interactions to gapped magnetic excitations. In practice, we adopt two different approaches: First, a model-based one relying on the Rashba Hamiltonian, which is employed to demystify and understand magnetic and transport properties of magnetic nanostructures embedded in a Rashba electron gas. Second, we use a first-principles approach within the framework of the Korringa-Kohn-Rostoker (KKR) Green function method to investigate the ground state properties of magnetic impurities in topologically insulating hosts. This method is suitable to simulate nanostructures in real space. Then, we employed our newly developed code based on time-dependent density functional theory to compute the spin excitation spectra of these magnetic nanostructures embedded in topological insulators. Moreover, the KKR Green function method was used to simulate the electronic structure and ground state properties of large magnetic nanostructures, namely magnetic Skyrmions. In the first part, the analytical Rashba Green function and the scattering matrices modeling the magnetic impurities in the s-wave approximation are employed for the computation of the magnetic interaction tensor which contains: isotropic exchange, Dzyaloshinskii-Moriya (DM) and pseudo-dipolar interactions. The competition between these interactions leads to a rich phase diagram depending on the distance between the magnetic impurities. Next, we consider an external perturbing electric field and investigate the transport properties by computing the residual resistivity tensor within linear response theory. The contribution of SOI is explored. The investigation of arbitrary orientations of the impurity magnetic moment allowed a detailed analysis of contributions from the anisotropic magnetoresistance and planar Hall effect. Moreover, we calculate the impurity induced bound currents in the Rashba electron gas, which are used to compute the induced orbital magnetization. For a trimer of impurities with a non-vanishing spin chirality (SC) a finite orbital magnetization is observed when SOI is turned off. Since it emerges from the SC, it was named chiral orbital magnetization. [...] Bouaziz, Juba},
isbn = {9783958064294},
langid = {english},
- keywords = {\_tablet,Hall QHE,Hall QSHE,juKKR,KKR,PGI-1/IAS-1,skyrmions,thesis,topological insulator},
+ keywords = {Hall QHE,Hall QSHE,juKKR,KKR,PGI-1/IAS-1,skyrmions,thesis,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Bouaziz_2019_Spin-orbitronics at the nanoscale.pdf;/Users/wasmer/Zotero/storage/YM28TKHA/865993.html}
}
@@ -1972,13 +2128,13 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Bran, Andres M. and Cox, Sam and Schilter, Oliver and Baldassari, Carlo and White, Andrew D. and Schwaller, Philippe},
date = {2023-10-02},
eprint = {2304.05376},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2304.05376},
url = {http://arxiv.org/abs/2304.05376},
urldate = {2023-10-08},
abstract = {Over the last decades, excellent computational chemistry tools have been developed. Integrating them into a single platform with enhanced accessibility could help reaching their full potential by overcoming steep learning curves. Recently, large-language models (LLMs) have shown strong performance in tasks across domains, but struggle with chemistry-related problems. Moreover, these models lack access to external knowledge sources, limiting their usefulness in scientific applications. In this study, we introduce ChemCrow, an LLM chemistry agent designed to accomplish tasks across organic synthesis, drug discovery, and materials design. By integrating 18 expert-designed tools, ChemCrow augments the LLM performance in chemistry, and new capabilities emerge. Our agent autonomously planned and executed the syntheses of an insect repellent, three organocatalysts, and guided the discovery of a novel chromophore. Our evaluation, including both LLM and expert assessments, demonstrates ChemCrow's effectiveness in automating a diverse set of chemical tasks. Surprisingly, we find that GPT-4 as an evaluator cannot distinguish between clearly wrong GPT-4 completions and Chemcrow's performance. Our work not only aids expert chemists and lowers barriers for non-experts, but also fosters scientific advancement by bridging the gap between experimental and computational chemistry.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AML,autonomous agent,chemistry,experimental,LangChain,library,LLM,ML,self-driving lab,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Bran et al_2023_ChemCrow.pdf;/Users/wasmer/Zotero/storage/FFH8F743/2304.html}
}
@@ -1988,7 +2144,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Brandstetter, Johannes and Welling, Max and Worrall, Daniel E.},
date = {2022-05-29},
eprint = {2202.07643},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
publisher = {arXiv},
doi = {10.48550/arXiv.2202.07643},
@@ -2043,13 +2199,13 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Brehmer, Paul and Herbst, Michael F. and Wessel, Stefan and Rizzi, Matteo and Stamm, Benjamin},
date = {2023-05-13},
eprint = {2304.13587},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:quant-ph},
doi = {10.48550/arXiv.2304.13587},
url = {http://arxiv.org/abs/2304.13587},
urldate = {2023-05-26},
abstract = {Within the reduced basis methods approach, an effective low-dimensional subspace of a quantum many-body Hilbert space is constructed in order to investigate, e.g., the ground-state phase diagram. The basis of this subspace is built from solutions of snapshots, i.e., ground states corresponding to particular and well-chosen parameter values. Here, we show how a greedy strategy to assemble the reduced basis and thus to select the parameter points can be implemented based on matrix-product-states (MPS) calculations. Once the reduced basis has been obtained, observables required for the computation of phase diagrams can be computed with a computational complexity independent of the underlying Hilbert space for any parameter value. We illustrate the efficiency and accuracy of this approach for different one-dimensional quantum spin-1 models, including anisotropic as well as biquadratic exchange interactions, leading to rich quantum phase diagrams.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Brehmer et al_2023_Reduced basis surrogates for quantum spin systems based on tensor networks.pdf;/Users/wasmer/Zotero/storage/KVDI7XMB/2304.html}
}
@@ -2075,6 +2231,22 @@ Subject\_term\_id: computational-chemistry;computational-science},
file = {/Users/wasmer/Nextcloud/Zotero/Breuck et al_2021_Robust model benchmarking and bias-imbalance in data-driven materials science.pdf}
}
+@thesis{brinkerComplexMagnetismNanostructures2021,
+ title = {Complex Magnetism of Nanostructures on Surfaces: From Orbital Magnetism to Spin Excitations},
+ shorttitle = {Complex Magnetism of Nanostructures on Surfaces},
+ author = {Brinker, Sascha},
+ date = {2021},
+ number = {FZJ-2021-01786},
+ institution = {Forschungszentrum Jülich GmbH Zentralbibliothek, Verlag},
+ url = {https://juser.fz-juelich.de/record/891866},
+ urldate = {2024-06-06},
+ abstract = {Magnetic nanostructures on surfaces are promising building blocks of future spintronics devices, as they represent the ultimate limit in miniaturization. In this thesis, a combination of density functional theory and model-based studies is used to investigate magnetic nanostructures on surfaces with respect to fundamental theoretical properties and in relation to scanning tunneling microscopy experiments. Novel properties are unveiled in this class of systems by several methodological developments, from a new perspective on the orbital magnetism to the static and dynamic properties of complex non-collinear magnetic states. Firstly, we shed light on the orbital magnetic moment in magnetic nanostructures on surfacesand find a new component – the inter-atomic orbital moment. A systematic analysis uncoversits distinct physical origin, its non-negligible strength, and its particular long range in realistic systems like adatoms deposited on the Pt(111) surface. Our results show unambiguously theimportance and the potential of this new contribution to the orbital magnetism.Secondly, we investigate magnetic exchange interactions in magnetic nanostructures goingbeyond the common bilinear exchange interactions. Special focus is given to higher-order interactions whose microscopic origin is clarified using a model-based study. Using the prototypical test systems of magnetic dimers we find a new chiral pair interaction, the chiral biquadratic interaction, which is the biquadratic equivalent to the well-known Dzyaloshinskii-Moriya interaction, and investigate its properties and its implications not only for finite nanostructures but also for extended systems. Thirdly, we focus on the spin dynamics and the damping in non-collinear magnetic structures by investigating the dependencies of the Gilbert damping tensor on the non-collinearity in an atomistic form using a combination of a model-based study and first-principles calculations. We show how isotropic and chiral dependencies evolve from an Anderson-like model and inrealistic systems like magnetic dimers on the Au(111) surface. These results have the potential to drive the field of atomistic spin dynamics to a more sophisticated description of the damping mechanisms. Fourthly, we investigate the magnetic stability of nanostructures, which is one of the key ingredients on the road towards future data storage devices. The impact of magnetic exchange interactions between nanostructures on the magnetic stability as probed in telegraph noise scanning tunneling microscopy experiments is analyzed by using the example of a magnetic trimer and a magnetic adatom. We find three regimes each driven by a distinct magnetic exchange interaction and show how this knowledge can be used to engineer the magnetic stability. Lastly, we analyze the complex interplay of magnetism, spin-orbit coupling and superconductivity in magnetic chains on a superconducting substrate with a special focus on the emergence of boundary states. We shed light on the puzzling magnetic ground state of Fe chains on theRe(0001) substrate and show how boundary effects can be minimized by termination with non-magnetic Co chains. Our results provide vital clues on the nature of the boundary states found in Fe chains on Re(0001), and support their identification as Majorana states. Brinker, Sascha},
+ isbn = {9783958065253},
+ langid = {english},
+ keywords = {/unread,exchange interaction,FZJ,juKKR,KKR,magnetism,PGI,PGI-1/IAS-1,surface physics,thesis},
+ file = {/Users/wasmer/Nextcloud/Zotero/Brinker_2021_Complex magnetism of nanostructures on surfaces.pdf}
+}
+
@article{brockherdeBypassingKohnShamEquations2017,
title = {Bypassing the {{Kohn-Sham}} Equations with Machine Learning},
author = {Brockherde, Felix and Vogt, Leslie and Li, Li and Tuckerman, Mark E. and Burke, Kieron and Müller, Klaus-Robert},
@@ -2092,7 +2264,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
abstract = {Last year, at least 30,000 scientific papers used the Kohn–Sham scheme of density functional theory to solve electronic structure problems in a wide variety of scientific fields. Machine learning holds the promise of learning the energy functional via examples, bypassing the need to solve the Kohn–Sham equations. This should yield substantial savings in computer time, allowing larger systems and/or longer time-scales to be tackled, but attempts to machine-learn this functional have been limited by the need to find its derivative. The present work overcomes this difficulty by directly learning the density-potential and energy-density maps for test systems and various molecules. We perform the first molecular dynamics simulation with a machine-learned density functional on malonaldehyde and are able to capture the intramolecular proton transfer process. Learning density models now allows the construction of accurate density functionals for realistic molecular systems.},
issue = {1},
langid = {english},
- keywords = {\_tablet,DFT,HK map,KRR,ML,ML-DFT,ML-ESM,ML-HK map,ML-KS,ML-OF,prediction from potential,prediction of electron density},
+ keywords = {DFT,HK map,KRR,ML,ML-DFT,ML-ESM,ML-HK map,ML-KS,ML-OF,prediction from potential,prediction of electron density},
file = {/Users/wasmer/Nextcloud/Zotero/Brockherde et al_2017_Bypassing the Kohn-Sham equations with machine learning.pdf;/Users/wasmer/Zotero/storage/8X4ALINZ/s41467-017-00839-3.html}
}
@@ -2151,7 +2323,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Bronstein, Michael M. and Bruna, Joan and Cohen, Taco and VeliÄković, Petar},
date = {2021-05-02},
eprint = {2104.13478},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
url = {http://arxiv.org/abs/2104.13478},
urldate = {2022-04-14},
@@ -2214,7 +2386,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
urldate = {2022-07-10},
abstract = {The purpose of this short essay is to introduce students and other newcomers to the basic ideas and uses of modern electronic density functional theory, including what kinds of approximations are in current use, and how well they work (or not). The complete newcomer should find it orients them well, while even longtime users and aficionados might find something new outside their area. Important questions varying in difficulty and effort are posed in the text, and are answered in the Supporting Information. © 2012 Wiley Periodicals, Inc.},
langid = {english},
- keywords = {\_tablet,density functional theory,electronic structure,local density approximation},
+ keywords = {density functional theory,electronic structure,local density approximation},
file = {/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Burke_Wagner_2013_DFT in a nutshell.pdf;/Users/wasmer/Zotero/storage/CCPHAAVK/qua.html}
}
@@ -2224,13 +2396,13 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Burke, Kieron and Kozlowski, John},
date = {2021-10-18},
eprint = {2108.11534},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2108.11534},
url = {http://arxiv.org/abs/2108.11534},
urldate = {2022-07-10},
abstract = {Most realistic calculations of moderately correlated materials begin with a ground-state density functional theory (DFT) calculation. While Kohn-Sham DFT is used in about 40,000 scientific papers each year, the fundamental underpinnings are not widely appreciated. In this chapter, we analyze the inherent characteristics of DFT in their simplest form, using the asymmetric Hubbard dimer as an illustrative model. We begin by working through the core tenets of DFT, explaining what the exact ground-state density functional yields and does not yield. Given the relative simplicity of the system, almost all properties of the exact exchange-correlation functional are readily visualized and plotted. Key concepts include the Kohn-Sham scheme, the behavior of the XC potential as correlations become very strong, the derivative discontinuity and the difference between KS gaps and true charge gaps, and how to extract optical excitations using time-dependent DFT. By the end of this text and accompanying exercises, the reader will improve their ability to both explain and visualize the concepts of DFT, as well as better understand where others may go wrong.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DFT,physics,strongly correlated maeterials},
file = {/Users/wasmer/Nextcloud/Zotero/Burke_Kozlowski_2021_Lies My Teacher Told Me About Density Functional Theory.pdf;/Users/wasmer/Zotero/storage/6EW6SVTP/2108.html}
}
@@ -2251,7 +2423,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
urldate = {2023-06-12},
abstract = {Data-driven methods based on machine learning have the potential to accelerate computational analysis of atomic structures. In this context, reliable uncertainty estimates are important for assessing confidence in predictions and enabling decision making. However, machine learning models can produce badly calibrated uncertainty estimates and it is therefore crucial to detect and handle uncertainty carefully. In this work we extend a message passing neural network designed specifically for predicting properties of molecules and materials with a calibrated probabilistic predictive distribution. The method presented in this paper differs from previous work by considering both aleatoric and epistemic uncertainty in a unified framework, and by recalibrating the predictive distribution on unseen data. Through computer experiments, we show that our approach results in accurate models for predicting molecular formation energies with well calibrated uncertainty in and out of the training data distribution on two public molecular benchmark datasets, QM9 and PC9. The proposed method provides a general framework for training and evaluating neural network ensemble models that are able to produce accurate predictions of properties of molecules with well calibrated uncertainty estimates.},
langid = {english},
- keywords = {\_tablet,active learning,AML,ensemble learning,HTC,ML,MLP,MPNN,PC9,QM9,uncertainty quantification},
+ keywords = {active learning,AML,ensemble learning,HTC,ML,MLP,MPNN,PC9,QM9,uncertainty quantification},
file = {/Users/wasmer/Nextcloud/Zotero/Busk et al_2021_Calibrated uncertainty for molecular property prediction using ensembles of.pdf}
}
@@ -2300,17 +2472,35 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Bystrom, Kyle and Kozinsky, Boris},
date = {2023-03-01},
eprint = {2303.00682},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2303.00682},
url = {http://arxiv.org/abs/2303.00682},
urldate = {2023-03-20},
abstract = {The design of better exchange-correlation (XC) functionals for Density Functional Theory (DFT) is a central challenge of modern electronic structure theory. However, current developments are limited by the mathematical form of the functional, with efficient semi-local functionals being inaccurate for many technologically important systems and the more accurate hybrid functionals being too expensive for large solid-state systems due to the use of the exact exchange operator. In this work, we use physics-informed machine learning to design an exchange functional that is both orbital-dependent and nonlocal, but which can be evaluated at roughly the cost of semi-local functionals and is significantly faster than hybrid DFT in plane-wave codes. By training functionals with several different feature sets, we elucidate the roles of orbital-dependent and nonlocal features in learning the exchange energy and determine that both types of features provide vital and independently important information to the model. Having trained our new exchange functional with an expressive, nonlocal feature set, we substitute it into existing hybrid functionals to achieve hybrid-DFT accuracy on thermochemical benchmark sets and improve the accuracy of band gap predictions over semi-local DFT. To demonstrate the scalability of our approach as well as the practical benefits of improved band gap prediction, we compute charged defect transition levels in silicon using large supercells. Due to its transferability and computational efficiency for both molecular and extended systems, our model overcomes the cost-accuracy trade-off between semi-local and hybrid DFT, and our general approach provides a feasible path toward a universal exchange-correlation functional with post-hybrid DFT accuracy and semi-local DFT cost.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {all-electron,AML,CIDER,DFT,ML,ML-DFA,ML-ESM,PAW,plane-wave,prediction of Exc},
file = {/Users/wasmer/Zotero/storage/FUN9D5UI/Bystrom and Kozinsky - 2023 - Nonlocal Machine-Learned Exchange Functional for M.pdf;/Users/wasmer/Zotero/storage/9YGJRHDG/2303.html}
}
+@article{cahenEnergyGlobalChallenge2008,
+ title = {Energy, the Global Challenge, and Materials},
+ author = {Cahen, David and Lubomirsky, Igor},
+ date = {2008-12-01},
+ journaltitle = {Materials Today},
+ shortjournal = {Materials Today},
+ volume = {11},
+ number = {12},
+ pages = {16--20},
+ issn = {1369-7021},
+ doi = {10.1016/S1369-7021(08)70248-7},
+ url = {https://www.sciencedirect.com/science/article/pii/S1369702108702487},
+ urldate = {2024-08-01},
+ abstract = {After some definitions to establish common ground and illustrate the issues in terms of orders of magnitude, we note that meeting the Energy challenge will require suitable materials. Luckily, we can count on the availability of natural resources for most materials. We briefly illustrate the connection between materials and energy and review the past and the present situations, to focus on the future. We wrap up by arguing that more than bare economics is required to use the fruits of science and technology towards a world order, built on sustainable energy (and materials) resources.},
+ keywords = {chemistry,energy challenge,energy materials,for introductions},
+ file = {/Users/wasmer/Nextcloud/Zotero/Cahen_Lubomirsky_2008_Energy, the global challenge, and materials.pdf;/Users/wasmer/Zotero/storage/C2JLRAGA/S1369702108702487.html}
+}
+
@article{caiSelfadaptiveFirstprinciplesApproach2023,
title = {A Self-Adaptive First-Principles Approach for Magnetic Excited States},
author = {Cai, Zefeng and Wang, Ke and Xu, Yong and Wei, Su-Huai and Xu, Ben},
@@ -2364,17 +2554,53 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Callow, Timothy J. and Kraisler, Eli and Cangi, Attila},
date = {2023-05-11},
eprint = {2305.06856},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2305.06856},
url = {http://arxiv.org/abs/2305.06856},
urldate = {2023-05-26},
abstract = {Rapid access to accurate equation-of-state (EOS) data is crucial in the warm-dense matter regime, as it is employed in various applications, such as providing input for hydrodynamics codes to model inertial confinement fusion processes. In this study, we develop neural network models for predicting the EOS based on first-principles data. The first model utilizes basic physical properties, while the second model incorporates more sophisticated physical information, using output from average-atom calculations as features. Average-atom models are often noted for providing a reasonable balance of accuracy and speed; however, our comparison of average-atom models and higher-fidelity calculations shows that more accurate models are required in the warm-dense matter regime. Both the neural network models we propose, particularly the physics-enhanced one, demonstrate significant potential as accurate and efficient methods for computing EOS data in warm-dense matter.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,EOS,todo-tagging,warm dense matter},
file = {/Users/wasmer/Nextcloud/Zotero/Callow et al_2023_Physics-enhanced neural networks for equation-of-state calculations.pdf;/Users/wasmer/Zotero/storage/W7JDGW6I/2305.html}
}
+@article{callowPhysicsenhancedNeuralNetworks2023a,
+ title = {Physics-Enhanced Neural Networks for Equation-of-State Calculations},
+ author = {Callow, Timothy J. and Nikl, Jan and Kraisler, Eli and Cangi, Attila},
+ date = {2023-12},
+ journaltitle = {Machine Learning: Science and Technology},
+ shortjournal = {Mach. Learn.: Sci. Technol.},
+ volume = {4},
+ number = {4},
+ pages = {045055},
+ publisher = {IOP Publishing},
+ issn = {2632-2153},
+ doi = {10.1088/2632-2153/ad13b9},
+ url = {https://dx.doi.org/10.1088/2632-2153/ad13b9},
+ urldate = {2024-06-07},
+ abstract = {Rapid access to accurate equation-of-state (EOS) data is crucial in the warm-dense matter (WDM) regime, as it is employed in various applications, such as providing input for hydrodynamic codes to model inertial confinement fusion processes. In this study, we develop neural network models for predicting the EOS based on first-principles data. The first model utilises basic physical properties, while the second model incorporates more sophisticated physical information, using output from average-atom (AA) calculations as features. AA models are often noted for providing a reasonable balance of accuracy and speed; however, our comparison of AA models and higher-fidelity calculations shows that more accurate models are required in the WDM regime. Both the neural network models we propose, particularly the physics-enhanced one, demonstrate significant potential as accurate and efficient methods for computing EOS data in WDM.},
+ langid = {english},
+ keywords = {/unread,AIMD,AML,average-atom model,cross-validation,EOS,feature engineering,ML,NN,physics-informed ML,warm dense matter},
+ file = {/Users/wasmer/Nextcloud/Zotero/Callow et al_2023_Physics-enhanced neural networks for equation-of-state calculations2.pdf}
+}
+
+@report{calvinIPCC2023Climate2023,
+ title = {{{IPCC}}, 2023: {{Climate Change}} 2023: {{Synthesis Report}}. {{Contribution}} of {{Working Groups I}}, {{II}} and {{III}} to the {{Sixth Assessment Report}} of the {{Intergovernmental Panel}} on {{Climate Change}} [{{Core Writing Team}}, {{H}}. {{Lee}} and {{J}}. {{Romero}} (Eds.)]. {{IPCC}}, {{Geneva}}, {{Switzerland}}.},
+ shorttitle = {{{IPCC}}, 2023},
+ author = {Calvin, Katherine and Dasgupta, Dipak and Krinner, Gerhard and Mukherji, Aditi and Thorne, Peter W. and Trisos, Christopher and Romero, José and Aldunce, Paulina and Barrett, Ko and Blanco, Gabriel and Cheung, William W.L. and Connors, Sarah and Denton, Fatima and Diongue-Niang, Aïda and Dodman, David and Garschagen, Matthias and Geden, Oliver and Hayward, Bronwyn and Jones, Christopher and Jotzo, Frank and Krug, Thelma and Lasco, Rodel and Lee, Yune-Yi and Masson-Delmotte, Valérie and Meinshausen, Malte and Mintenbeck, Katja and Mokssit, Abdalah and Otto, Friederike E.L. and Pathak, Minal and Pirani, Anna and Poloczanska, Elvira and Pörtner, Hans-Otto and Revi, Aromar and Roberts, Debra C. and Roy, Joyashree and Ruane, Alex C. and Skea, Jim and Shukla, Priyadarshi R. and Slade, Raphael and Slangen, Aimée and Sokona, Youba and Sörensson, Anna A. and Tignor, Melinda and Van Vuuren, Detlef and Wei, Yi-Ming and Winkler, Harald and Zhai, Panmao and Zommers, Zinta and Hourcade, Jean-Charles and Johnson, Francis X. and Pachauri, Shonali and Simpson, Nicholas P. and Singh, Chandni and Thomas, Adelle and Totin, Edmond and Arias, Paola and Bustamante, Mercedes and Elgizouli, Ismail and Flato, Gregory and Howden, Mark and Méndez-Vallejo, Carlos and Pereira, Joy Jacqueline and Pichs-Madruga, Ramón and Rose, Steven K. and Saheb, Yamina and Sánchez RodrÃguez, Roberto and Ãœrge-Vorsatz, Diana and Xiao, Cunde and Yassaa, Noureddine and AlegrÃa, Andrés and Armour, Kyle and Bednar-Friedl, Birgit and Blok, Kornelis and Cissé, Guéladio and Dentener, Frank and Eriksen, Siri and Fischer, Erich and Garner, Gregory and Guivarch, Céline and Haasnoot, Marjolijn and Hansen, Gerrit and Hauser, Mathias and Hawkins, Ed and Hermans, Tim and Kopp, Robert and Leprince-Ringuet, Noëmie and Lewis, Jared and Ley, Debora and Ludden, Chloé and Niamir, Leila and Nicholls, Zebedee and Some, Shreya and Szopa, Sophie and Trewin, Blair and Van Der Wijst, Kaj-Ivar and Winter, Gundula and Witting, Maximilian and Birt, Arlene and Ha, Meeyoung and Romero, José and Kim, Jinmi and Haites, Erik F. and Jung, Yonghun and Stavins, Robert and Birt, Arlene and Ha, Meeyoung and Orendain, Dan Jezreel A. and Ignon, Lance and Park, Semin and Park, Youngin and Reisinger, Andy and Cammaramo, Diego and Fischlin, Andreas and Fuglestvedt, Jan S. and Hansen, Gerrit and Ludden, Chloé and Masson-Delmotte, Valérie and Matthews, J.B. Robin and Mintenbeck, Katja and Pirani, Anna and Poloczanska, Elvira and Leprince-Ringuet, Noëmie and Péan, Clotilde},
+ namea = {Lee, Hoesung},
+ nameatype = {collaborator},
+ date = {2023-07-25},
+ edition = {First},
+ institution = {Intergovernmental Panel on Climate Change (IPCC)},
+ doi = {10.59327/IPCC/AR6-9789291691647},
+ url = {https://www.ipcc.ch/report/ar6/syr/},
+ urldate = {2024-08-02},
+ abstract = {The Synthesis Report (SYR) is a stand-alone synthesis of the most policy-relevant evidence from the scientific, technical, and socio-economic literature assessed in the Sixth Assessment Report (AR6) of the Intergovernmental Panel on Climate Change (IPCC). The SYR distils and integrates the main findings of the three reports of the Working Groups of the IPCC during the AR6, and the three AR6 Special Reports into a concise document. It consists of a Summary for Policymakers and a longer report.},
+ keywords = {/unread}
+}
+
@book{cancesDensityFunctionalTheory2023,
title = {Density {{Functional Theory}}: {{Modeling}}, {{Mathematical Analysis}}, {{Computational Methods}}, and {{Applications}}},
shorttitle = {Density {{Functional Theory}}},
@@ -2399,7 +2625,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
volume = {113},
number = {1},
eprint = {2210.04512},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
pages = {21},
issn = {0377-9017, 1573-0530},
@@ -2425,7 +2651,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
url = {https://link.aps.org/doi/10.1103/PhysRevA.88.062505},
urldate = {2022-07-08},
abstract = {Potential functional approximations are an intriguing alternative to density functional approximations. The potential functional that is dual to the Lieb density functional is defined and its properties are reported. The relationship between the Thomas-Fermi theory as a density functional and the theory as a potential functional is derived. The properties of several recent semiclassical potential functionals are explored, especially regarding their approach to the large particle number and classical continuum limits. The lack of ambiguity in the energy density of potential functional approximations is demonstrated. The density-density response function of the semiclassical approximation is calculated and shown to violate a key symmetry condition.},
- keywords = {\_tablet,density functional,density vs potential,DFT,potential},
+ keywords = {density functional,density vs potential,DFT,potential},
file = {/Users/wasmer/Nextcloud/Zotero/Cangi et al_2013_Potential functionals versus density functionals.pdf;/Users/wasmer/Zotero/storage/4U87YYPT/Cangi et al_2013_Potential functionals versus density functionals.pdf;/Users/wasmer/Zotero/storage/AJH43GTS/PhysRevA.88.html}
}
@@ -2495,11 +2721,11 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Capelle, Klaus},
date = {2006-11-18},
eprint = {cond-mat/0211443},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
url = {http://arxiv.org/abs/cond-mat/0211443},
urldate = {2021-08-31},
abstract = {This paper is the outgrowth of lectures the author gave at the Physics Institute and the Chemistry Institute of the University of Sao Paulo at Sao Carlos, Brazil, and at the VIII'th Summer School on Electronic Structure of the Brazilian Physical Society. It is an attempt to introduce density-functional theory (DFT) in a language accessible for students entering the field or researchers from other fields. It is not meant to be a scholarly review of DFT, but rather an informal guide to its conceptual basis and some recent developments and advances. The Hohenberg-Kohn theorem and the Kohn-Sham equations are discussed in some detail. Approximate density functionals, selected aspects of applications of DFT, and a variety of extensions of standard DFT are also discussed, albeit in less detail. Throughout it is attempted to provide a balanced treatment of aspects that are relevant for chemistry and aspects relevant for physics, but with a strong bias towards conceptual foundations. The paper is intended to be read before (or in parallel with) one of the many excellent more technical reviews available in the literature.},
- keywords = {\_tablet,DFT,learn DFT,review},
+ keywords = {DFT,DFT theory,educational,electronic structure theory,learn DFT,learning material,review},
file = {/Users/wasmer/Nextcloud/Zotero/Capelle_2006_A bird's-eye view of density-functional theory.pdf;/Users/wasmer/Zotero/storage/8TLEU4M3/0211443.html}
}
@@ -2508,13 +2734,13 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Carbone, Johanna P. and Bouaziz, Juba and Bihlmayer, Gustav and Blügel, Stefan},
date = {2023-09-27},
eprint = {2309.15513},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2309.15513},
url = {http://arxiv.org/abs/2309.15513},
urldate = {2023-10-04},
abstract = {Rare-earth (RE) atoms on top of 2D materials represent an interesting platform with the prospect of tailoring the magnetic anisotropy for practical applications. Here, we investigate the ground state and magnetic properties of selected \$4f\$-atoms deposited on a graphene substrate in the framework of the DFT+\$U\$ approach. The inherent strong spin-orbit interaction in conjunction with crystal field effects acting on the localized \$4f\$-shells results in a substantial magnetic anisotropy energy (tens of meVs), whose angular dependence is dictated by the \$C\_\{6v\}\$ symmetry of the graphene substrate. We obtain the crystal field parameters and investigate spin-flip events via quantum tunneling of magnetization in the view of achieving a protected quantum-spin behavior. Remarkably, the large spin and orbital moments of the open \$4f\$-shells (Dy, Ho and Tm) generate a strong magneto-elastic coupling which provides more flexibility to control the magnetic state via the application of external strain.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {2D material,adatoms,DFT,DFT+U,FLEUR,FZJ,magnetic anisotropy,magnetism,PGI,PGI-1/IAS-1,physics,quantum materials,rare earths,surface physics},
file = {/Users/wasmer/Nextcloud/Zotero/Carbone et al_2023_Investigation of magnetic properties of $4f$-adatoms on graphene.pdf;/Users/wasmer/Zotero/storage/RWZ9JMI5/2309.html}
}
@@ -2562,7 +2788,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
url = {https://link.aps.org/doi/10.1103/PhysRevB.100.024112},
urldate = {2021-05-13},
abstract = {We explore different ways to simplify the evaluation of the smooth overlap of atomic positions (SOAP) many-body atomic descriptor [Bartók et al., Phys. Rev. B 87, 184115 (2013).]. Our aim is to improve the computational efficiency of SOAP-based similarity kernel construction. While these improved atomic descriptors can be used for general characterization and interpolation of atomic properties, their main target application is accelerated evaluation of machine-learning-based interatomic potentials within the Gaussian approximation potential (GAP) framework [Bartók et al., Phys. Rev. Lett. 104, 136403 (2010)]. We achieve this objective by expressing the atomic densities in an approximate separable form, which decouples the radial and angular channels. We then express the elements of the SOAP descriptor (i.e., the expansion coefficients for the atomic densities) in analytical form given a particular choice of radial basis set. Finally, we derive recursion formulas for the expansion coefficients. This new SOAP-based descriptor allows for tenfold speedups compared to previous implementations, while improving the stability of the radial expansion for distant atomic neighbors, without degradation of the interpolation power of GAP models.},
- keywords = {\_tablet,descriptor dimred,descriptors,descriptors analysis,dimensionality reduction,GAP,ML,MLP,SOAP},
+ keywords = {descriptor dimred,descriptors,descriptors analysis,dimensionality reduction,GAP,ML,MLP,SOAP},
file = {/Users/wasmer/Nextcloud/Zotero/Caro_2019_Optimizing many-body atomic descriptors for enhanced computational performance.pdf;/Users/wasmer/Zotero/storage/FDHHHJTR/PhysRevB.100.html}
}
@@ -2661,14 +2887,14 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Casares, Pablo A. M. and Baker, Jack S. and Medvidovic, Matija and family=Reis, given=Roberto, prefix=dos, useprefix=false and Arrazola, Juan Miguel},
date = {2023-09-22},
eprint = {2309.15127},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, physics:quant-ph},
doi = {10.48550/arXiv.2309.15127},
url = {http://arxiv.org/abs/2309.15127},
urldate = {2023-10-05},
abstract = {Density functional theory (DFT) stands as a cornerstone method in computational quantum chemistry and materials science due to its remarkable versatility and scalability. Yet, it suffers from limitations in accuracy, particularly when dealing with strongly correlated systems. To address these shortcomings, recent work has begun to explore how machine learning can expand the capabilities of DFT; an endeavor with many open questions and technical challenges. In this work, we present Grad DFT: a fully differentiable JAX-based DFT library, enabling quick prototyping and experimentation with machine learning-enhanced exchange-correlation energy functionals. Grad DFT employs a pioneering parametrization of exchange-correlation functionals constructed using a weighted sum of energy densities, where the weights are determined using neural networks. Moreover, Grad DFT encompasses a comprehensive suite of auxiliary functions, notably featuring a just-in-time compilable and fully differentiable self-consistent iterative procedure. To support training and benchmarking efforts, we additionally compile a curated dataset of experimental dissociation energies of dimers, half of which contain transition metal atoms characterized by strong electronic correlations. The software library is tested against experimental results to study the generalization capabilities of a neural functional across potential energy surfaces and atomic species, as well as the effect of training data noise on the resulting model accuracy.},
- pubstate = {preprint},
- keywords = {/unread,\_tablet,AML,autodiff,DM21,JAX,library,ML,ML-DFA,ML-DFT,ML-ESM,prediction of Exc,transition metals,with-code},
+ pubstate = {prepublished},
+ keywords = {/unread,AML,autodiff,DM21,JAX,library,ML,ML-DFA,ML-DFT,ML-ESM,prediction of Exc,transition metals,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Casares et al_2023_Grad DFT.pdf;/Users/wasmer/Zotero/storage/EZ4L7B7D/2309.html}
}
@@ -2722,7 +2948,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
urldate = {2022-12-29},
abstract = {Over the past decade, interatomic potentials based on machine~learning (ML) techniques have become an indispensable tool in the atomic-scale modeling of materials. Trained on energies and forces obtained from electronic-structure calculations, they inherit their predictive accuracy, and extend greatly the length and time scales that are accessible to explicit atomistic simulations. Inexpensive predictions of the energetics of individual configurations have facilitated greatly the calculation of the thermodynamics of materials, including finite-temperature effects and disorder. More recently, ML models have been closing the gap with first-principles calculations in another area: the prediction of arbitrarily complicated functional properties, from vibrational and optical spectroscopies to electronic excitations. The implementation of integrated ML models that combine energetic and functional predictions with statistical and dynamical sampling of atomic-scale properties is bringing the promise of predictive, uncompromising simulations of existing and novel materials closer to its full realization.},
langid = {english},
- keywords = {\_tablet,equivariant,Gibbs free energy,integrated models,MLP,multiscale,prediction of DOS,prediction of polarizability,review,symmetry,tensorial target,thermodynamics},
+ keywords = {equivariant,Gibbs free energy,integrated models,MLP,multiscale,prediction of DOS,prediction of polarizability,review,symmetry,tensorial target,thermodynamics},
file = {/Users/wasmer/Nextcloud/Zotero/Ceriotti_2022_Beyond potentials.pdf}
}
@@ -2772,14 +2998,14 @@ Subject\_term\_id: computational-chemistry;computational-science},
author = {Cersonsky, Rose K. and Pakhnova, Maria and Engel, Edgar A. and Ceriotti, Michele},
date = {2022-12-22},
eprint = {2209.10709},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, stat},
doi = {10.48550/arXiv.2209.10709},
url = {http://arxiv.org/abs/2209.10709},
urldate = {2023-01-23},
abstract = {Due to the subtle balance of intermolecular interactions that govern structure-property relations, predicting the stability of crystal structures formed from molecular building blocks is a highly non-trivial scientific problem. A particularly active and fruitful approach involves classifying the different combinations of interacting chemical moieties, as understanding the relative energetics of different interactions enables the design of molecular crystals and fine-tuning their stabilities. While this is usually performed based on the empirical observation of the most commonly encountered motifs in known crystal structures, we propose to apply a combination of supervised and unsupervised machine-learning techniques to automate the construction of an extensive library of molecular building blocks. We introduce a structural descriptor tailored to the prediction of the binding (lattice) energy and apply it to a curated dataset of organic crystals and exploit its atom-centered nature to obtain a data-driven assessment of the contribution of different chemical groups to the lattice energy of the crystal. We then interpret this library using a low-dimensional representation of the structure-energy landscape and discuss selected examples of the insights into crystal engineering that can be extracted from this analysis, providing a complete database to guide the design of molecular materials.},
- pubstate = {preprint},
- keywords = {\_tablet,ACDC,crystal structure,dimensionality reduction,linear regression,organic crystals,PCovR,SOAP,structure prediction,structure search,unsupervised learning},
+ pubstate = {prepublished},
+ keywords = {ACDC,crystal structure,dimensionality reduction,linear regression,organic crystals,PCovR,SOAP,structure prediction,structure search,unsupervised learning},
file = {/Users/wasmer/Nextcloud/Zotero/Cersonsky et al_2022_A data-driven interpretation of the stability of molecular crystals.pdf;/Users/wasmer/Zotero/storage/25LFMGQ5/2209.html}
}
@@ -2799,7 +3025,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
urldate = {2022-08-10},
abstract = {Selecting the most relevant features and samples out of a large set of candidates is a task that occurs very often in the context of automated data analysis, where it improves the computational performance and often the transferability of a model. Here we focus on two popular subselection schemes applied to this end: CUR decomposition, derived from a low-rank approximation of the feature matrix, and farthest point sampling (FPS), which relies on the iterative identification of the most diverse samples and discriminating features. We modify these unsupervised approaches, incorporating a supervised component following the same spirit as the principal covariates (PCov) regression method. We show how this results in selections that perform better in supervised tasks, demonstrating with models of increasing complexity, from ridge regression to kernel ridge regression and finally feed-forward neural networks. We also present adjustments to minimise the impact of any subselection when performing unsupervised tasks. We demonstrate the significant improvements associated with PCov-CUR and PCov-FPS selections for applications to chemistry and materials science, typically reducing by a factor of two the number of features and samples required to achieve a given level of regression accuracy.},
langid = {english},
- keywords = {\_tablet,CUR decomposition,dimensionality reduction,feature selection,FPS,KPCovR,KRR,PCovR,sample selection},
+ keywords = {CUR decomposition,dimensionality reduction,feature selection,FPS,KPCovR,KRR,PCovR,sample selection},
file = {/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Cersonsky et al_2021_Improving sample and feature selection with principal covariates regression.pdf}
}
@@ -2833,7 +3059,7 @@ Subject\_term\_id: computational-chemistry;computational-science},
abstract = {Simulations based on solving the Kohn-Sham (KS) equation of density functional theory (DFT) have become a vital component of modern materials and chemical sciences research and development portfolios. Despite its versatility, routine DFT calculations are usually limited to a few hundred atoms due to the computational bottleneck posed by the KS equation. Here we introduce a machine-learning-based scheme to efficiently assimilate the function of the KS equation, and by-pass it to directly, rapidly, and accurately predict the electronic structure of a material or a molecule, given just its atomic configuration. A new rotationally invariant representation is utilized to map the atomic environment around a grid-point to the electron density and local density of states at that grid-point. This mapping is learned using a neural network trained on previously generated reference DFT results at millions of grid-points. The proposed paradigm allows for the high-fidelity emulation of KS DFT, but orders of magnitude faster than the direct solution. Moreover, the machine learning prediction scheme is strictly linear-scaling with system size.},
issue = {1},
langid = {english},
- keywords = {\_tablet,custom structural descriptors,descriptors,DFT,FCNN,grid-based descriptors,LDOS,ML,ML-DFT,ML-ESM,models,NN,prediction from structure,prediction of electron density,prediction of LDOS,RNN},
+ keywords = {custom structural descriptors,descriptors,DFT,FCNN,grid-based descriptors,LDOS,ML,ML-DFT,ML-ESM,models,NN,prediction from structure,prediction of electron density,prediction of LDOS,RNN},
annotation = {Bandiera\_abtest: a\\
Cc\_license\_type: cc\_by\\
Cg\_type: Nature Research Journals\\
@@ -2843,6 +3069,25 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
file = {/Users/wasmer/Nextcloud/Zotero/Chandrasekaran et al_2019_Solving the electronic structure problem with machine learning.pdf;/Users/wasmer/Nextcloud/Zotero/Chandrasekaran et al_2019_Solving the electronic structure problem with machine learning2.pdf;/Users/wasmer/Zotero/storage/TL92B668/s41524-019-0162-7.html}
}
+@article{changColloquiumQuantumAnomalous2023,
+ title = {Colloquium: {{Quantum}} Anomalous {{Hall}} Effect},
+ shorttitle = {Colloquium},
+ author = {Chang, Cui-Zu and Liu, Chao-Xing and MacDonald, Allan H.},
+ date = {2023-01-23},
+ journaltitle = {Reviews of Modern Physics},
+ shortjournal = {Rev. Mod. Phys.},
+ volume = {95},
+ number = {1},
+ pages = {011002},
+ publisher = {American Physical Society},
+ doi = {10.1103/RevModPhys.95.011002},
+ url = {https://link.aps.org/doi/10.1103/RevModPhys.95.011002},
+ urldate = {2024-07-05},
+ abstract = {The quantum Hall (QH) effect, quantized Hall resistance combined with zero longitudinal resistance, is the characteristic experimental fingerprint of Chern insulators—topologically nontrivial states of two-dimensional matter with broken time-reversal symmetry. In Chern insulators, nontrivial bulk band topology is expressed by chiral states that carry current along sample edges without dissipation. The quantum anomalous Hall (QAH) effect refers to QH effects that occur in the absence of external magnetic fields due to spontaneously broken time-reversal symmetry. The QAH effect has now been realized in four different classes of two-dimensional materials: (i) thin films of magnetically (Cr- and/or V-) doped topological insulators in the (Bi,Sb)2â¢Te3 family, (ii) thin films of the intrinsic magnetic topological insulator MnBi2â¢Te4, (iii) moiré materials formed from graphene, and (iv) moiré materials formed from transition-metal dichalcogenides. In this Colloquium, the physical mechanisms responsible for each class of QAH insulator are reviewed, with both differences and commonalities highlighted, and potential applications of the QAH effect are commented upon.},
+ keywords = {/unread,colloquium,condensed matter,for introductions,Hall effect,Hall QAHE,physics,topological,topological insulator},
+ file = {/Users/wasmer/Nextcloud/Zotero/Chang et al_2023_Colloquium.pdf;/Users/wasmer/Zotero/storage/LZ2CCRTE/RevModPhys.95.html}
+}
+
@article{changExperimentalObservationQuantum2013,
title = {Experimental {{Observation}} of the {{Quantum Anomalous Hall Effect}} in a {{Magnetic Topological Insulator}}},
author = {Chang, Cui-Zu and Zhang, Jinsong and Feng, Xiao and Shen, Jie and Zhang, Zuocheng and Guo, Minghua and Li, Kang and Ou, Yunbo and Wei, Pang and Wang, Li-Li and Ji, Zhong-Qing and Feng, Yang and Ji, Shuaihua and Chen, Xi and Jia, Jinfeng and Dai, Xi and Fang, Zhong and Zhang, Shou-Cheng and He, Ke and Wang, Yayu and Lu, Li and Ma, Xu-Cun and Xue, Qi-Kun},
@@ -2855,6 +3100,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
doi = {10.1126/science.1234414},
url = {https://www.science.org/doi/10.1126/science.1234414},
urldate = {2022-05-13},
+ keywords = {experimental,for introductions,groundbreaking,Hall effect,Hall QAHE,magnetic topological materials,magnetism,topological,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Chang et al_2013_Experimental Observation of the Quantum Anomalous Hall Effect in a Magnetic.pdf}
}
@@ -2903,7 +3149,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
volume = {11},
number = {10},
eprint = {2010.09990},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
pages = {6059--6072},
issn = {2155-5435, 2155-5435},
@@ -2921,7 +3167,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Chard, Ryan and Li, Zhuozhao and Chard, Kyle and Ward, Logan and Babuji, Yadu and Woodard, Anna and Tuecke, Steve and Blaiszik, Ben and Franklin, Michael J. and Foster, Ian},
date = {2018-11-27},
eprint = {1811.11213},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
url = {http://arxiv.org/abs/1811.11213},
urldate = {2022-01-03},
@@ -2930,9 +3176,25 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
file = {/Users/wasmer/Nextcloud/Zotero/Chard et al_2018_DLHub.pdf;/Users/wasmer/Zotero/storage/VT5H6PP6/1811.html}
}
+@online{chengEquivariantNeuralOperator2023,
+ title = {Equivariant {{Neural Operator Learning}} with {{Graphon Convolution}}},
+ author = {Cheng, Chaoran and Peng, Jian},
+ date = {2023-11-17},
+ eprint = {2311.10908},
+ eprinttype = {arXiv},
+ eprintclass = {cs},
+ doi = {10.48550/arXiv.2311.10908},
+ url = {http://arxiv.org/abs/2311.10908},
+ urldate = {2024-05-28},
+ abstract = {We propose a general architecture that combines the coefficient learning scheme with a residual operator layer for learning mappings between continuous functions in the 3D Euclidean space. Our proposed model is guaranteed to achieve SE(3)-equivariance by design. From the graph spectrum view, our method can be interpreted as convolution on graphons (dense graphs with infinitely many nodes), which we term InfGCN. By leveraging both the continuous graphon structure and the discrete graph structure of the input data, our model can effectively capture the geometric information while preserving equivariance. Through extensive experiments on large-scale electron density datasets, we observed that our model significantly outperformed the current state-of-the-art architectures. Multiple ablation studies were also carried out to demonstrate the effectiveness of the proposed architecture.},
+ pubstate = {prepublished},
+ keywords = {AML,equivariant,FNO,graphon convolution,ML,ML-Density,ML-DFT,ML-ESM,neural operator,prediction of electron density,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Cheng_Peng_2023_Equivariant Neural Operator Learning with Graphon Convolution.pdf;/Users/wasmer/Zotero/storage/WJENKT52/2311.html}
+}
+
@article{chengMappingMaterialsMolecules2020,
title = {Mapping {{Materials}} and {{Molecules}}},
- author = {Cheng, Bingqing and Griffiths, Ryan-Rhys and Wengert, Simon and Kunkel, Christian and Stenczel, Tamas and Zhu, Bonan and Deringer, Volker L. and Bernstein, Noam and Margraf, Johannes T. and Reuter, Karsten and Csanyi, Gabor},
+ author = {Cheng, Bingqing and Griffiths, Ryan-Rhys and Wengert, Simon and Kunkel, Christian and Stenczel, Tamas and Zhu, Bonan and Deringer, Volker L. and Bernstein, Noam and Margraf, Johannes T. and Reuter, Karsten and Csányi, Gábor},
date = {2020-09-15},
journaltitle = {Accounts of Chemical Research},
shortjournal = {Acc. Chem. Res.},
@@ -2964,7 +3226,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
url = {https://doi.org/10.1021/acs.chemmater.9b01294},
urldate = {2022-01-02},
abstract = {Graph networks are a new machine learning (ML) paradigm that supports both relational reasoning and combinatorial generalization. Here, we develop universal MatErials Graph Network (MEGNet) models for accurate property prediction in both molecules and crystals. We demonstrate that the MEGNet models outperform prior ML models such as the SchNet in 11 out of 13 properties of the QM9 molecule data set. Similarly, we show that MEGNet models trained on ∼60 000 crystals in the Materials Project substantially outperform prior ML models in the prediction of the formation energies, band gaps, and elastic moduli of crystals, achieving better than density functional theory accuracy over a much larger data set. We present two new strategies to address data limitations common in materials science and chemistry. First, we demonstrate a physically intuitive approach to unify four separate molecular MEGNet models for the internal energy at 0 K and room temperature, enthalpy, and Gibbs free energy into a single free energy MEGNet model by incorporating the temperature, pressure, and entropy as global state inputs. Second, we show that the learned element embeddings in MEGNet models encode periodic chemical trends and can be transfer-learned from a property model trained on a larger data set (formation energies) to improve property models with smaller amounts of data (band gaps and elastic moduli).},
- keywords = {\_tablet,GNN,library,MEGNet,molecules,solids,vectorial learning target,with-code},
+ keywords = {AML,crystal graph,GNN,library,materials,MEGNet,ML,MLP,molecules,original publication,solids,vectorial learning target,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Chen et al_2019_Graph Networks as a Universal Machine Learning Framework for Molecules and.pdf}
}
@@ -2978,7 +3240,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
urldate = {2023-03-20},
abstract = {Machine learning (ML) has been widely applied to chemical property prediction, most prominently for the energies and forces in molecules and materials. The strong interest in predicting energies in particular has led to a local energy-based paradigm for modern atomistic ML models, which ensures size-extensivity and a linear scaling of computational cost with system size. However, many electronic properties (such as excitation energies or ionization energies) do not necessarily scale linearly with system size and may even be spatially localized. Using size-extensive models in these cases can lead to large errors. In this work, we explore different strategies for learning intensive and localized properties, using HOMO energies in organic molecules as a representative test case. In particular, we analyze the pooling functions that atomistic neural networks use to predict molecular properties, and suggest an orbital weighted average (OWA) approach that enables the accurate prediction of orbital energies and locations.},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,HOMO,intensive properties,localized electronic structure,ML,model comparison,MPNN,pooling,prediction of intensive properties,prediction of orbital energies,SchNet,SchNetPack,SOAP},
file = {/Users/wasmer/Zotero/storage/HAJCMMQ8/Chen et al. - 2023 - Physics-Inspired Machine Learning of Localized Int.pdf}
}
@@ -2988,15 +3250,52 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Chen, Chi and Ong, Shyue Ping},
date = {2022-02-04},
eprint = {2202.02450},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2202.02450},
urldate = {2022-03-28},
abstract = {Interatomic potentials (IAPs), which describe the potential energy surface of a collection of atoms, are a fundamental input for atomistic simulations. However, existing IAPs are either fitted to narrow chemistries or too inaccurate for general applications. Here, we report a universal IAP for materials based on graph neural networks with three-body interactions (M3GNet). The M3GNet IAP was trained on the massive database of structural relaxations performed by the Materials Project over the past 10 years and has broad applications in structural relaxation, dynamic simulations and property prediction of materials across diverse chemical spaces. About 1.8 million potentially stable materials were identified from a screening of 31 million hypothetical crystal structures, demonstrating a machine learning-accelerated pathway to the discovery of synthesizable materials with exceptional properties.},
- keywords = {\_tablet,condensed matter,GNN,library,M3GNet,materials,materials database,materials project,matterverse,MEGNet,ML,MLP,molecules,periodic table,solids,tensorial target,universal potential,with-code},
+ keywords = {condensed matter,GNN,library,M3GNet,materials,materials database,materials project,matterverse,MEGNet,ML,MLP,molecules,original publication,periodic table,solids,tensorial target,universal potential,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Chen_Ong_2022_A Universal Graph Deep Learning Interatomic Potential for the Periodic Table.pdf;/Users/wasmer/Zotero/storage/H4FKVKUF/2202.html}
}
+@article{chenUniversalGraphDeep2022a,
+ title = {A Universal Graph Deep Learning Interatomic Potential for the Periodic Table},
+ author = {Chen, Chi and Ong, Shyue Ping},
+ date = {2022-11},
+ journaltitle = {Nature Computational Science},
+ shortjournal = {Nat Comput Sci},
+ volume = {2},
+ number = {11},
+ pages = {718--728},
+ publisher = {Nature Publishing Group},
+ issn = {2662-8457},
+ doi = {10.1038/s43588-022-00349-3},
+ url = {https://www.nature.com/articles/s43588-022-00349-3},
+ urldate = {2024-05-23},
+ abstract = {Interatomic potentials (IAPs), which describe the potential energy surface of atoms, are a fundamental input for atomistic simulations. However, existing IAPs are either fitted to narrow chemistries or too inaccurate for general applications. Here we report a universal IAP for materials based on graph neural networks with three-body interactions (M3GNet). The M3GNet IAP was trained on the massive database of structural relaxations performed by the Materials Project over the past ten years and has broad applications in structural relaxation, dynamic simulations and property prediction of materials across diverse chemical spaces. About 1.8 million materials from a screening of 31 million hypothetical crystal structures were identified to be potentially stable against existing Materials Project crystals based on M3GNet energies. Of the top 2,000 materials with the lowest energies above the convex hull, 1,578 were verified to be stable using density functional theory calculations. These results demonstrate a machine learning-accelerated pathway to the discovery of synthesizable materials with exceptional properties.},
+ langid = {english},
+ keywords = {/unread,condensed matter,GNN,library,M3GNet,materials,materials database,materials project,matterverse,MEGNet,ML,MLP,molecules,original publication,periodic table,solids,tensorial target,universal potential,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Chen_Ong_2022_A universal graph deep learning interatomic potential for the periodic table3.pdf}
+}
+
+@online{chiangLLaMPLargeLanguage2024,
+ title = {{{LLaMP}}: {{Large Language Model Made Powerful}} for {{High-fidelity Materials Knowledge Retrieval}} and {{Distillation}}},
+ shorttitle = {{{LLaMP}}},
+ author = {Chiang, Yuan and Hsieh, Elvis and Chou, Chia-Hong and Riebesell, Janosh},
+ date = {2024-06-02},
+ eprint = {2401.17244},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat},
+ doi = {10.48550/arXiv.2401.17244},
+ url = {http://arxiv.org/abs/2401.17244},
+ urldate = {2024-06-13},
+ abstract = {Reducing hallucination of Large Language Models (LLMs) is imperative for use in the sciences, where reliability and reproducibility are crucial. However, LLMs inherently lack long-term memory, making it a nontrivial, ad hoc, and inevitably biased task to fine-tune them on domain-specific literature and data. Here we introduce LLaMP, a multimodal retrieval-augmented generation (RAG) framework of hierarchical reasoning-and-acting (ReAct) agents that can dynamically and recursively interact with computational and experimental data on Materials Project (MP) and run atomistic simulations via high-throughput workflow interface. Without fine-tuning, LLaMP demonstrates strong tool usage ability to comprehend and integrate various modalities of materials science concepts, fetch relevant data stores on the fly, process higher-order data (such as crystal structure and elastic tensor), and streamline complex tasks in computational materials and chemistry. We propose a simple metric combining uncertainty and confidence estimates to evaluate the self-consistency of responses by LLaMP and vanilla LLMs. Our benchmark shows that LLaMP effectively mitigates the intrinsic bias in LLMs, counteracting the errors on bulk moduli, electronic bandgaps, and formation energies that seem to derive from mixed data sources. We also demonstrate LLaMP's capability to edit crystal structures and run annealing molecular dynamics simulations using pre-trained machine-learning force fields. The framework offers an intuitive and nearly hallucination-free approach to exploring and scaling materials informatics, and establishes a pathway for knowledge distillation and fine-tuning other language models. Code and live demo are available at https://github.com/chiang-yuan/llamp},
+ pubstate = {prepublished},
+ keywords = {/unread,AML,chemical synthesis,fine-tuning,hallucinations,intelligent agent,language models,language-driven simulation,LLM,materials,materials project,MD,ML,MLP,multi-agent system,multi-modal,prediction of magnetic order,retrieval-augmented generation,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Chiang et al_2024_LLaMP.pdf;/Users/wasmer/Zotero/storage/UJZQMJHQ/2401.html}
+}
+
@article{chngMachineLearningPhases2017,
title = {Machine {{Learning Phases}} of {{Strongly Correlated Fermions}}},
author = {Ch’ng, Kelvin and Carrasquilla, Juan and Melko, Roger G. and Khatami, Ehsan},
@@ -3020,17 +3319,32 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Chong, Sanggyu and Grasselli, Federico and Mahmoud, Chiheb Ben and Morrow, Joe D. and Deringer, Volker L. and Ceriotti, Michele},
date = {2023-06-27},
eprint = {2306.15638},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2306.15638},
url = {http://arxiv.org/abs/2306.15638},
urldate = {2023-07-01},
abstract = {Machine learning (ML) models for molecules and materials commonly rely on a decomposition of the global target quantity into local, atom-centered contributions. This approach is convenient from a computational perspective, enabling large-scale ML-driven simulations with a linear-scaling cost, and also allow for the identification and post-hoc interpretation of contributions from individual chemical environments and motifs to complicated macroscopic properties. However, even though there exist practical justifications for these decompositions, only the global quantity is rigorously defined, and thus it is unclear to what extent the atomistic terms predicted by the model can be trusted. Here, we introduce a quantitative metric, which we call the local prediction rigidity (LPR), that allows one to assess how robust the locally decomposed predictions of ML models are. We investigate the dependence of LPR on the aspects of model training, particularly the composition of training dataset, for a range of different problems from simple toy models to real chemical systems. We present strategies to systematically enhance the LPR, which can be used to improve the robustness, interpretability, and transferability of atomistic ML models.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Chong et al_2023_Robustness of Local Predictions in Atomistic Machine Learning Models2.pdf;/Users/wasmer/Zotero/storage/SEXVHR6B/2306.html}
}
+@book{choudharyArtificialIntelligenceScience2023,
+ title = {Artificial {{Intelligence}} for {{Science}}: {{A Deep Learning Revolution}}},
+ shorttitle = {Artificial {{Intelligence}} for {{Science}}},
+ author = {Choudhary, Alok and Fox, Geoffrey and Hey, Tony},
+ date = {2023-04},
+ publisher = {WORLD SCIENTIFIC},
+ doi = {10.1142/13123},
+ url = {https://www.worldscientific.com/worldscibooks/10.1142/13123},
+ urldate = {2024-05-16},
+ abstract = {This unique collection introduces AI, Machine Learning (ML), and deep neural network technologies leading to scientific discovery from the datasets generated both by supercomputer simulation and by modern experimental facilities. Huge quantities of experimental data come from many sources — telescopes, satellites, gene sequencers, accelerators, and electron microscopes, including international facilities such as the Large Hadron Collider (LHC) at CERN in Geneva and the ITER Tokamak in France. These sources generate many petabytes moving to exabytes of data per year. Extracting scientific insights from these data is a major challenge for scientists, for whom the latest AI developments will be essential. The timely handbook benefits professionals, researchers, academics, and students in all fields of science and engineering as well as AI, ML, and neural networks. Further, the vision evident in this book inspires all those who influence or are influenced by scientific progress.},
+ isbn = {9789811265662 9789811265679},
+ langid = {english},
+ keywords = {/unread,AI4Science,AlphaFold,AML,artificial intelligence,for introductions,ML,review}
+}
+
@article{choudharyAtomisticLineGraph2021,
title = {Atomistic {{Line Graph Neural Network}} for Improved Materials Property Predictions},
author = {Choudhary, Kamal and DeCost, Brian},
@@ -3052,6 +3366,27 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
file = {/Users/wasmer/Nextcloud/Zotero/Choudhary_DeCost_2021_Atomistic Line Graph Neural Network for improved materials property predictions.pdf;/Users/wasmer/Zotero/storage/F8XSYTPV/s41524-021-00650-1.html}
}
+@article{choudharyJARVISLeaderboardLargeScale2024,
+ title = {{{JARVIS-Leaderboard}}: A Large Scale Benchmark of Materials Design Methods},
+ shorttitle = {{{JARVIS-Leaderboard}}},
+ author = {Choudhary, Kamal and Wines, Daniel and Li, Kangming and Garrity, Kevin F. and Gupta, Vishu and Romero, Aldo H. and Krogel, Jaron T. and Saritas, Kayahan and Fuhr, Addis and Ganesh, Panchapakesan and Kent, Paul R. C. and Yan, Keqiang and Lin, Yuchao and Ji, Shuiwang and Blaiszik, Ben and Reiser, Patrick and Friederich, Pascal and Agrawal, Ankit and Tiwary, Pratyush and Beyerle, Eric and Minch, Peter and Rhone, Trevor David and Takeuchi, Ichiro and Wexler, Robert B. and Mannodi-Kanakkithodi, Arun and Ertekin, Elif and Mishra, Avanish and Mathew, Nithin and Wood, Mitchell and Rohskopf, Andrew Dale and Hattrick-Simpers, Jason and Wang, Shih-Han and Achenie, Luke E. K. and Xin, Hongliang and Williams, Maureen and Biacchi, Adam J. and Tavazza, Francesca},
+ date = {2024-05-07},
+ journaltitle = {npj Computational Materials},
+ shortjournal = {npj Comput Mater},
+ volume = {10},
+ number = {1},
+ pages = {1--17},
+ publisher = {Nature Publishing Group},
+ issn = {2057-3960},
+ doi = {10.1038/s41524-024-01259-w},
+ url = {https://www.nature.com/articles/s41524-024-01259-w},
+ urldate = {2024-07-23},
+ abstract = {Lack of rigorous reproducibility and validation are significant hurdles for scientific development across many fields. Materials science, in particular, encompasses a variety of experimental and theoretical approaches that require careful benchmarking. Leaderboard efforts have been developed previously to mitigate these issues. However, a comprehensive comparison and benchmarking on an integrated platform with multiple data modalities with perfect and defect materials data is still lacking. This work introduces JARVIS-Leaderboard, an open-source and community-driven platform that facilitates benchmarking and enhances reproducibility. The platform allows users to set up benchmarks with custom tasks and enables contributions in the form of dataset, code, and meta-data submissions. We cover the following materials design categories: Artificial Intelligence (AI), Electronic Structure (ES), Force-fields (FF), Quantum Computation (QC), and Experiments (EXP). For AI, we cover several types of input data, including atomic structures, atomistic images, spectra, and text. For ES, we consider multiple ES approaches, software packages, pseudopotentials, materials, and properties, comparing results to experiment. For FF, we compare multiple approaches for material property predictions. For QC, we benchmark Hamiltonian simulations using various quantum algorithms and circuits. Finally, for experiments, we use the inter-laboratory approach to establish benchmarks. There are 1281 contributions to 274 benchmarks using 152 methods with more than 8 million data points, and the leaderboard is continuously expanding. The JARVIS-Leaderboard is available at the website: https://pages.nist.gov/jarvis\_leaderboard/},
+ langid = {english},
+ keywords = {/unread,AML,benchmarking,Database,dataset,DFT,educational,JARVIS,JARVIS-DFT,learning material,ML,ML-FF,ML-IAP,model comparison,model reporting,model repository,RDM,reproducibility,RSE,scientific workflows,version control,with-code,with-data},
+ file = {/Users/wasmer/Nextcloud/Zotero/Choudhary et al_2024_JARVIS-Leaderboard.pdf}
+}
+
@online{choudharyLargeScaleBenchmark2023,
title = {Large {{Scale Benchmark}} of {{Materials Design Methods}}},
author = {Choudhary, Kamal and Wines, Daniel and Li, Kangming and Garrity, Kevin F. and Gupta, Vishu and Romero, Aldo H. and Krogel, Jaron T. and Saritas, Kayahan and Fuhr, Addis and Ganesh, Panchapakesan and Kent, Paul R. C. and Yan, Keqiang and Lin, Yuchao and Ji, Shuiwang and Blaiszik, Ben and Reiser, Patrick and Friederich, Pascal and Agrawal, Ankit and Tiwary, Pratyush and Beyerle, Eric and Minch, Peter and Rhone, Trevor David and Takeuchi, Ichiro and Wexler, Robert B. and Mannodi-Kanakkithodi, Arun and Ertekin, Elif and Mishra, Avanish and Mathew, Nithin and Baird, Sterling G. and Wood, Mitchell and Rohskopf, Andrew Dale and Hattrick-Simpers, Jason and Wang, Shih-Han and Achenie, Luke E. K. and Xin, Hongliang and Williams, Maureen and Biacchi, Adam J. and Tavazza, Francesca},
@@ -3060,7 +3395,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
urldate = {2023-07-01},
abstract = {Lack of rigorous reproducibility and validation are major hurdles for scientific development across many fields. Materials science in particular encompasses a variety of experimental and theoretical approaches that require careful benchmarking. Leaderboard efforts have been developed previously to mitigate these issues. However, a comprehensive comparison and benchmarking on an integrated platform with multiple data modalities with both perfect and defect materials data is still lacking. This work introduces JARVIS-Leaderboard, an open-source and community-driven platform that facilitates benchmarking and enhances reproducibility. The platform allows users to set up benchmarks with custom tasks and enables contributions in the form of dataset, code, and meta-data submissions. We cover the following materials design categories: Artificial Intelligence (AI), Electronic Structure (ES), Force-fields (FF), Quantum Computation (QC) and Experiments (EXP). For AI, we cover several types of input data, including atomic structures, atomistic images, spectra, and text. For ES, we consider multiple ES approaches, software packages, pseudopotentials, materials, and properties, comparing results to experiment. For FF, we compare multiple approaches for material property predictions. For QC, we benchmark Hamiltonian simulations using various quantum algorithms and circuits. Finally, for experiments, we use the inter-laboratory approach to establish benchmarks. There are 1281 contributions to 274 benchmarks using 152 methods with more than 8 million data-points, and the leaderboard is continuously expanding. The JARVIS-Leaderboard is available at the website: https://pages.nist.gov/jarvis\_leaderboard},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {benchmarking,JARVIS,materials database,todo-tagging,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Choudhary et al_2023_Large Scale Benchmark of Materials Design Methods.pdf}
}
@@ -3089,7 +3424,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Chouhan, Prashant Singh and Khetawat, Harsh and Resnik, Neil and Jain, Twinkle and Garg, Rohan and Cooperman, Gene and Hartman-Baker, Rebecca and Zhao, Zhengji},
date = {2021-03-16},
eprint = {2103.08546},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
url = {http://arxiv.org/abs/2103.08546},
urldate = {2021-10-20},
@@ -3123,13 +3458,13 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Cignoni, Edoardo and Suman, Divya and Nigam, Jigyasa and Cupellini, Lorenzo and Mennucci, Benedetta and Ceriotti, Michele},
date = {2023-11-01},
eprint = {2311.00844},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2311.00844},
url = {http://arxiv.org/abs/2311.00844},
urldate = {2023-11-05},
abstract = {Data-driven techniques are increasingly used to replace electronic-structure calculations of matter. In this context, a relevant question is whether machine learning (ML) should be applied directly to predict the desired properties or be combined explicitly with physically-grounded operations. We present an example of an integrated modeling approach, in which a symmetry-adapted ML model of an effective Hamiltonian is trained to reproduce electronic excitations from a quantum-mechanical calculation. The resulting model can make predictions for molecules that are much larger and more complex than those that it is trained on, and allows for dramatic computational savings by indirectly targeting the outputs of well-converged calculations while using a parameterization corresponding to a minimal atom-centered basis. These results emphasize the merits of intertwining data-driven techniques with physical approximations, improving the transferability and interpretability of ML models without affecting their accuracy and computational efficiency, and providing a blueprint for developing ML-augmented electronic-structure methods.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,B3LYP,basis set,emulator,excited states,hybrid AI/simulation,library,ML,ML-DFT,ML-ESM,ML-WFT,molecular orbitals,molecules,partial charges,prediction of charge transfer,prediction of Hamiltonian matrix,prediction of orbital energies,STO-3G,transfer learning,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Cignoni et al_2023_Electronic excited states from physically-constrained machine learning.pdf;/Users/wasmer/Zotero/storage/XWXG8UVG/2311.html}
}
@@ -3175,7 +3510,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Cobelli, Matteo and Cahalane, Paddy and Sanvito, Stefano},
date = {2022-01-27},
eprint = {2201.11591},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
url = {http://arxiv.org/abs/2201.11591},
urldate = {2022-03-23},
@@ -3251,7 +3586,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
urldate = {2023-06-30},
abstract = {We reproduced the simulations described in Wang et al. (2018) and found we could not obtain the results reported. The root cause was identified to be incorrect atom masses in the original simulation files. As a consequence, the potential does not reproduce the experimental glass density – and presumably, other structural properties – and should be considered with great caution.},
langid = {english},
- keywords = {/unread,AML,glasses,ML,MLP,reproduction study,skepticism},
+ keywords = {AML,glasses,ML,MLP,reproduction study,skepticism},
file = {/Users/wasmer/Nextcloud/Zotero/Coudert_2023_Failure to reproduce the results of “A new transferable interatomic potential.pdf;/Users/wasmer/Zotero/storage/GIWCGKS3/S0022309323002892.html}
}
@@ -3281,7 +3616,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
volume = {113},
number = {2},
eprint = {2207.05794},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {math-ph, physics:quant-ph},
pages = {42},
issn = {1573-0530},
@@ -3354,13 +3689,13 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Curcic, Milan},
date = {2019-03-25},
eprint = {1902.06714},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.1902.06714},
url = {http://arxiv.org/abs/1902.06714},
urldate = {2023-09-30},
abstract = {This paper describes neural-fortran, a parallel Fortran framework for neural networks and deep learning. It features a simple interface to construct feed-forward neural networks of arbitrary structure and size, several activation functions, and stochastic gradient descent as the default optimization algorithm. Neural-fortran also leverages the Fortran 2018 standard collective subroutines to achieve data-based parallelism on shared- or distributed-memory machines. First, I describe the implementation of neural networks with Fortran derived types, whole-array arithmetic, and collective sum and broadcast operations to achieve parallelism. Second, I demonstrate the use of neural-fortran in an example of recognizing hand-written digits from images. Finally, I evaluate the computational performance in both serial and parallel modes. Ease of use and computational performance are similar to an existing popular machine learning framework, making neural-fortran a viable candidate for further development and use in production.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,alternative approaches,Deep learning,Fortran,General ML,HPC,library,ML,NN,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Curcic_2019_A parallel Fortran framework for neural networks and deep learning.pdf;/Users/wasmer/Zotero/storage/PINUPP44/1902.html}
}
@@ -3420,7 +3755,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
abstract = {High-throughput computational approaches combining thermodynamic and electronic-structure methods with data mining and database construction are increasingly used to analyse huge amounts of data for the discovery and design of new materials. This Review provides an overall perspective of the field for a broad range of materials, and discusses upcoming challenges and opportunities.},
issue = {3},
langid = {english},
- keywords = {AFLOWLIB,Automation,compositional descriptors,Database,descriptor comparison,descriptors,DFT,High-throughput,HTC,materials database,materials discovery,materials screening,review},
+ keywords = {\_tablet,AFLOWLIB,Automation,compositional descriptors,Database,descriptor comparison,descriptors,DFT,High-throughput,HTC,materials database,materials discovery,materials screening,review},
file = {/Users/wasmer/Nextcloud/Zotero/Curtarolo et al_2013_The high-throughput highway to computational materials design.pdf}
}
@@ -3448,14 +3783,14 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Damewood, James and Karaguesian, Jessica and Lunger, Jaclyn R. and Tan, Aik Rui and Xie, Mingrou and Peng, Jiayu and Gómez-Bombarelli, Rafael},
date = {2023-01-20},
eprint = {2301.08813},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2301.08813},
url = {http://arxiv.org/abs/2301.08813},
urldate = {2023-03-19},
abstract = {High-throughput data generation methods and machine learning (ML) algorithms have given rise to a new era of computational materials science by learning relationships among composition, structure, and properties and by exploiting such relations for design. However, to build these connections, materials data must be translated into a numerical form, called a representation, that can be processed by a machine learning model. Datasets in materials science vary in format (ranging from images to spectra), size, and fidelity. Predictive models vary in scope and property of interests. Here, we review context-dependent strategies for constructing representations that enable the use of materials as inputs or outputs of machine learning models. Furthermore, we discuss how modern ML techniques can learn representations from data and transfer chemical and physical information between tasks. Finally, we outline high-impact questions that have not been fully resolved and thus, require further investigation.},
- pubstate = {preprint},
- keywords = {\_tablet,AML,defects,descriptors,disordered,materials,ML,review,review-of-descriptors,TODO},
+ pubstate = {prepublished},
+ keywords = {AML,defect engineering,defects,descriptors,disordered,materials,ML,review,review-of-descriptors,TODO},
file = {/Users/wasmer/Nextcloud/Zotero/Damewood et al_2023_Representations of Materials for Machine Learning.pdf;/Users/wasmer/Zotero/storage/7JZ95596/2301.html}
}
@@ -3465,13 +3800,13 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Danka, Tivadar and Horvath, Peter},
date = {2018-12-12},
eprint = {1805.00979},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.1805.00979},
url = {http://arxiv.org/abs/1805.00979},
urldate = {2023-04-10},
abstract = {modAL is a modular active learning framework for Python, aimed to make active learning research and practice simpler. Its distinguishing features are (i) clear and modular object oriented design (ii) full compatibility with scikit-learn models and workflows. These features make fast prototyping and easy extensibility possible, aiding the development of real-life active learning pipelines and novel algorithms as well. modAL is fully open source, hosted on GitHub at https://github.com/cosmic-cortex/modAL. To assure code quality, extensive unit tests are provided and continuous integration is applied. In addition, a detailed documentation with several tutorials are also available for ease of use. The framework is available in PyPI and distributed under the MIT license.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {active learning,General ML,library,Python},
file = {/Users/wasmer/Nextcloud/Zotero/Danka_Horvath_2018_modAL.pdf;/Users/wasmer/Zotero/storage/BRMUZWH3/1805.html}
}
@@ -3481,12 +3816,12 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Darby, James P. and Kermode, James R. and Csányi, Gábor},
date = {2021-12-24},
eprint = {2112.13055},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
url = {http://arxiv.org/abs/2112.13055},
urldate = {2022-01-03},
abstract = {Many atomic descriptors are currently limited by their unfavourable scaling with the number of chemical elements \$S\$ e.g. the length of body-ordered descriptors, such as the Smooth Overlap of Atomic Positions (SOAP) power spectrum (3-body) and the Atomic Cluster Expansion (ACE) (multiple body-orders), scales as \$(NS)\textasciicircum\textbackslash nu\$ where \$\textbackslash nu+1\$ is the body-order and \$N\$ is the number of radial basis functions used in the density expansion. We introduce two distinct approaches which can be used to overcome this scaling for the SOAP power spectrum. Firstly, we show that the power spectrum is amenable to lossless compression with respect to both \$S\$ and \$N\$, so that the descriptor length can be reduced from \$\textbackslash mathcal\{O\}(N\textasciicircum 2S\textasciicircum 2)\$ to \$\textbackslash mathcal\{O\}\textbackslash left(NS\textbackslash right)\$. Secondly, we introduce a generalized SOAP kernel, where compression is achieved through the use of the total, element agnostic density, in combination with radial projection. The ideas used in the generalized kernel are equally applicably to any other body-ordered descriptors and we demonstrate this for the Atom Centered Symmetry Functions (ACSF). Finally, both compression approaches are shown to offer comparable performance to the original descriptor across a variety of numerical tests.},
- keywords = {\_tablet,ACE,ACSF,chemical species scaling problem,descriptor dimred,descriptors,descriptors analysis,dimensionality reduction,library,ML,SOAP,with-code},
+ keywords = {ACE,ACSF,chemical species scaling problem,descriptor dimred,descriptors,descriptors analysis,dimensionality reduction,library,ML,SOAP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Darby et al_2021_Compressing local atomic neighbourhood descriptors.pdf;/Users/wasmer/Zotero/storage/GXXQQPAA/2112.html}
}
@@ -3507,7 +3842,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
abstract = {Many atomic descriptors are currently limited by their unfavourable scaling with the number of chemical elements S e.g. the length of body-ordered descriptors, such as the SOAP power spectrum (3-body) and the (ACE) (multiple body-orders), scales as (NS)ν where ν\,+\,1 is the body-order and N is the number of radial basis functions used in the density expansion. We introduce two distinct approaches which can be used to overcome this scaling for the SOAP power spectrum. Firstly, we show that the power spectrum is amenable to lossless compression with respect to both S and N, so that the descriptor length can be reduced from \$\$\{\{\{\textbackslash mathcal\{O\}\}\}\}(\{N\}\textasciicircum\{2\}\{S\}\textasciicircum\{2\})\$\$to \$\$\{\{\{\textbackslash mathcal\{O\}\}\}\}\textbackslash left(NS\textbackslash right)\$\$. Secondly, we introduce a generalised SOAP kernel, where compression is achieved through the use of the total, element agnostic density, in combination with radial projection. The ideas used in the generalised kernel are equally applicably to any other body-ordered descriptors and we demonstrate this for the (ACSF).},
issue = {1},
langid = {english},
- keywords = {\_tablet,ACE,ACSF,chemical species scaling problem,descriptor dimred,descriptors,descriptors analysis,dimensionality reduction,HEA,library,ML,SOAP,with-code},
+ keywords = {ACE,ACSF,chemical species scaling problem,descriptor dimred,descriptors,descriptors analysis,dimensionality reduction,HEA,library,ML,SOAP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Darby et al_2022_Compressing local atomic neighbourhood descriptors.pdf;/Users/wasmer/Zotero/storage/WR6IJ7MC/s41524-022-00847-y.html}
}
@@ -3519,7 +3854,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
url = {https://arxiv.org/abs/2210.01705},
urldate = {2023-12-18},
abstract = {Density based representations of atomic environments that are invariant under Euclidean symmetries have become a widely used tool in the machine learning of interatomic potentials, broader data-driven atomistic modelling and the visualisation and analysis of materials datasets.The standard mechanism used to incorporate chemical element information is to create separate densities for each element and form tensor products between them. This leads to a steep scaling in the size of the representation as the number of elements increases. Graph neural networks, which do not explicitly use density representations, escape this scaling by mapping the chemical element information into a fixed dimensional space in a learnable way. We recast this approach as tensor factorisation by exploiting the tensor structure of standard neighbour density based descriptors. In doing so, we form compact tensor-reduced representations whose size does not depend on the number of chemical elements, but remain systematically convergeable and are therefore applicable to a wide range of data analysis and regression tasks.},
- pubstate = {preprint},
+ pubstate = {prepublished},
version = {2},
keywords = {ACE,AML,chemical species scaling problem,descriptor comparison,descriptor dimred,descriptors,descriptors analysis,dimensionality reduction,HEA,invariance,MACE,ML,MPNN,Multi-ACE,organic chemistry,SOAP,todo-tagging,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Darby et al_2022_Tensor-reduced atomic density representations2.pdf}
@@ -3615,6 +3950,24 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
file = {/Users/wasmer/Nextcloud/Zotero/Davidson_Freire_2008_Provenance and scientific workflows.pdf}
}
+@article{daviesComputationalScreeningAll2016,
+ title = {Computational {{Screening}} of {{All Stoichiometric Inorganic Materials}}},
+ author = {Davies, Daniel W. and Butler, Keith T. and Jackson, Adam J. and Morris, Andrew and Frost, Jarvist M. and Skelton, Jonathan M. and Walsh, Aron},
+ date = {2016-10-13},
+ journaltitle = {Chem},
+ shortjournal = {Chem},
+ volume = {1},
+ number = {4},
+ pages = {617--627},
+ issn = {2451-9294},
+ doi = {10.1016/j.chempr.2016.09.010},
+ url = {https://www.sciencedirect.com/science/article/pii/S2451929416301553},
+ urldate = {2024-08-02},
+ abstract = {Forming a four-component compound from the first 103 elements of the periodic table results in more than 1012 combinations. Such a materials space is intractable to high-throughput experiment or first-principle computation. We~introduce a framework to address this problem and quantify how many materials can exist. We apply principles of valency and electronegativity to filter chemically implausible compositions, which reduces the inorganic quaternary space to 1010 combinations. We demonstrate that estimates of band gaps and absolute electron energies can be made simply on the basis of the chemical composition and apply this to the search for new semiconducting materials to support the photoelectrochemical splitting of water. We show the applicability to predicting crystal structure by analogy with known compounds, including exploration of the phase space for ternary combinations that form a perovskite lattice. Computer screening reproduces known perovskite materials and predicts the feasibility of thousands more. Given the simplicity of the approach, large-scale searches can be performed on a single workstation.},
+ keywords = {alloys,chemical space,DFT,HTC,materials,materials discovery,materials screening,perovskites,quaternary systems},
+ file = {/Users/wasmer/Nextcloud/Zotero/Davies et al_2016_Computational Screening of All Stoichiometric Inorganic Materials.pdf;/Users/wasmer/Zotero/storage/ECXMNEAF/S2451929416301553.html}
+}
+
@unpublished{dawidModernApplicationsMachine2022,
title = {Modern Applications of Machine Learning in Quantum Sciences},
author = {Dawid, Anna and Arnold, Julian and Requena, Borja and Gresch, Alexander and Płodzień, Marcin and Donatella, Kaelan and Nicoli, Kim A. and Stornati, Paolo and Koch, Rouven and Büttner, Miriam and Okuła, Robert and Muñoz-Gil, Gorka and Vargas-Hernández, Rodrigo A. and Cervera-Lierta, Alba and Carrasquilla, Juan and Dunjko, Vedran and Gabrié, Marylou and Huembeli, Patrick and family=Nieuwenburg, given=Evert, prefix=van, useprefix=true and Vicentini, Filippo and Wang, Lei and Wetzel, Sebastian J. and Carleo, Giuseppe and Greplová, Eliška and Krems, Roman and Marquardt, Florian and Tomza, Michał and Lewenstein, Maciej and Dauphin, Alexandre},
@@ -3632,13 +3985,13 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Dawid, Anna and Arnold, Julian and Requena, Borja and Gresch, Alexander and Płodzień, Marcin and Donatella, Kaelan and Nicoli, Kim A. and Stornati, Paolo and Koch, Rouven and Büttner, Miriam and Okuła, Robert and Muñoz-Gil, Gorka and Vargas-Hernández, Rodrigo A. and Cervera-Lierta, Alba and Carrasquilla, Juan and Dunjko, Vedran and Gabrié, Marylou and Huembeli, Patrick and family=Nieuwenburg, given=Evert, prefix=van, useprefix=true and Vicentini, Filippo and Wang, Lei and Wetzel, Sebastian J. and Carleo, Giuseppe and Greplová, Eliška and Krems, Roman and Marquardt, Florian and Tomza, Michał and Lewenstein, Maciej and Dauphin, Alexandre},
date = {2023-11-15},
eprint = {2204.04198},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:quant-ph},
doi = {10.48550/arXiv.2204.04198},
url = {http://arxiv.org/abs/2204.04198},
urldate = {2024-02-28},
abstract = {In this book, we provide a comprehensive introduction to the most recent advances in the application of machine learning methods in quantum sciences. We cover the use of deep learning and kernel methods in supervised, unsupervised, and reinforcement learning algorithms for phase classification, representation of many-body quantum states, quantum feedback control, and quantum circuits optimization. Moreover, we introduce and discuss more specialized topics such as differentiable programming, generative models, statistical approach to machine learning, and quantum machine learning.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,condensed matter,educational,equivariant,general ML,Ising,learning material,lecture notes,magnetic structure,ML,ML-QMBP,ML-WFT,NQS,prediction of wavefunction,quantum science,quantum state tomography,spin,spin symmetry,spin texture,symmetrization,symmetry,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Dawid et al_2023_Modern applications of machine learning in quantum sciences.pdf;/Users/wasmer/Zotero/storage/RCZ83DUM/2204.html}
}
@@ -3696,6 +4049,24 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
file = {/Users/wasmer/Nextcloud/Zotero/De et al_2016_Comparing molecules and solids across structural and alchemical space.pdf;/Users/wasmer/Zotero/storage/SA8QCH28/C6CP00415F.html}
}
+@incollection{dederichsKorringaKohnRostokerKKRGreen2006,
+ title = {The {{Korringa-Kohn-Rostoker}} ({{KKR}}) {{Green Function Method II}}. {{Impurities}} and {{Clusters}} in the {{Bulk}} and on {{Surfaces}}},
+ booktitle = {Computational {{Nanoscience}}: {{Do It Yourself}}! - {{Lecture Notes}}},
+ author = {Dederichs, P. H. and Zeller, Rudolf and Lounis, Samir},
+ date = {2006},
+ series = {{{NIC}} Series},
+ volume = {31},
+ publisher = {John von Neumann Institute for Computing},
+ location = {Jülich},
+ url = {http://hdl.handle.net/2128/4793},
+ urldate = {2024-07-04},
+ abstract = {Dederichs, P. H.; Lounis, S.; Zeller, R.},
+ isbn = {3-00-017350-1},
+ langid = {english},
+ keywords = {all-electron,defects,DFT,DFT theory,educational,Electronic structure,electronic structure theory,full-potential,full-relativistic,FZJ,impurity embedding,juKKR,KKR,KKR foundations,Korringa–Kohn–Rostoker,learning material,library,PGI-1/IAS-1,surface physics,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Dederichs et al_2006_The Korringa-Kohn-Rostoker (KKR) Green Function Method II.pdf;/Users/wasmer/Zotero/storage/KW5A8NAB/51355.html}
+}
+
@article{degraveMagneticControlTokamak2022,
title = {Magnetic Control of Tokamak Plasmas through Deep Reinforcement Learning},
author = {Degrave, Jonas and Felici, Federico and Buchli, Jonas and Neunert, Michael and Tracey, Brendan and Carpanese, Francesco and Ewalds, Timo and Hafner, Roland and Abdolmaleki, Abbas and family=Casas, given=Diego, prefix=de las, useprefix=true and Donner, Craig and Fritz, Leslie and Galperti, Cristian and Huber, Andrea and Keeling, James and Tsimpoukelli, Maria and Kay, Jackie and Merle, Antoine and Moret, Jean-Marc and Noury, Seb and Pesamosca, Federico and Pfau, David and Sauter, Olivier and Sommariva, Cristian and Coda, Stefano and Duval, Basil and Fasoli, Ambrogio and Kohli, Pushmeet and Kavukcuoglu, Koray and Hassabis, Demis and Riedmiller, Martin},
@@ -3727,6 +4098,25 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
keywords = {/unread,educational,General ML,learning material,linear algebra,mathematics,ML,ML theory,online book,probability theory,statistics,textbook}
}
+@article{delgado-liconaResearchAccelerationSelfDriving2023,
+ title = {Research {{Acceleration}} in {{Self-Driving Labs}}: {{Technological Roadmap}} toward {{Accelerated Materials}} and {{Molecular Discovery}}},
+ shorttitle = {Research {{Acceleration}} in {{Self-Driving Labs}}},
+ author = {Delgado-Licona, Fernando and Abolhasani, Milad},
+ date = {2023},
+ journaltitle = {Advanced Intelligent Systems},
+ volume = {5},
+ number = {4},
+ pages = {2200331},
+ issn = {2640-4567},
+ doi = {10.1002/aisy.202200331},
+ url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/aisy.202200331},
+ urldate = {2024-08-01},
+ abstract = {The urgency of finding solutions to global energy, sustainability, and healthcare challenges has motivated rethinking of the conventional chemistry and material science workflows. Self-driving labs, emerged through integration of disruptive physical and digital technologies, including robotics, additive manufacturing, reaction miniaturization, and artificial intelligence, have the potential to accelerate the pace of materials and molecular discovery by 10–100X. Using autonomous robotic experimentation workflows, self-driving labs enable access to a larger part of the chemical universe and reduce the time-to-solution through an iterative hypothesis formulation, intelligent experiment selection, and automated testing. By providing a data-centric abstraction to the accelerated discovery cycle, in this perspective article, the required hardware and software technological infrastructure to unlock the true potential of self-driving labs is discussed. In particular, process intensification as an accelerator mechanism for reaction modules of self-driving labs and digitalization strategies to further accelerate the discovery cycle in chemical and materials sciences are discussed.},
+ langid = {english},
+ keywords = {accelerated discovery,autonomous experimentation,autonomous research systems,digital labs,for introductions,lab automation,materials acceleration platforms,materials discovery,process intensification,self-driving lab,self-driving labs},
+ file = {/Users/wasmer/Nextcloud/Zotero/Delgado-Licona_Abolhasani_2023_Research Acceleration in Self-Driving Labs.pdf}
+}
+
@article{delrioDeepLearningFramework2023,
title = {A Deep Learning Framework to Emulate Density Functional Theory},
author = {family=Rio, given=Beatriz G., prefix=del, useprefix=true and Phan, Brandon and Ramprasad, Rampi},
@@ -3744,7 +4134,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
abstract = {Density functional theory (DFT) has been a critical component of computational materials research and discovery for decades. However, the computational cost of solving the central Kohn–Sham equation remains a major obstacle for dynamical studies of complex phenomena at-scale. Here, we propose an end-to-end machine learning (ML) model that emulates the essence of DFT by mapping the atomic structure of the system to its electronic charge density, followed by the prediction of other properties such as density of states, potential energy, atomic forces, and stress tensor, by using the atomic structure and charge density as input. Our deep learning model successfully bypasses the explicit solution of the Kohn-Sham equation with orders of magnitude speedup (linear scaling with system size with a small prefactor), while maintaining chemical accuracy. We demonstrate the capability of this ML-DFT concept for an extensive database of organic molecules, polymer chains, and polymer crystals.},
issue = {1},
langid = {english},
- keywords = {\_tablet,ACDC,AGNI desriptor,all-electron,AML,chemistry,Database,descriptors,emulator,grid-based descriptors,invariance,library,linear scaling,materials,ML,ML-DFT,ML-ESM,molecules,multi-output learning,multi-step model,NN,organic chemistry,PBE,polymers,prediction of bandgap,prediction of DOS,prediction of electron density,prediction of energy,prediction of forces,TensorFlow,tensorial target,VASP,with-code,with-data},
+ keywords = {ACDC,AGNI desriptor,all-electron,AML,chemistry,Database,descriptors,emulator,grid-based descriptors,invariance,library,linear scaling,materials,ML,ML-DFT,ML-ESM,molecules,multi-output learning,multi-step model,NN,organic chemistry,PBE,polymers,prediction of bandgap,prediction of DOS,prediction of electron density,prediction of energy,prediction of forces,TensorFlow,tensorial target,VASP,with-code,with-data},
file = {/Users/wasmer/Zotero/storage/EHN4XYXG/del Rio et al_2023_A deep learning framework to emulate density functional theory.pdf}
}
@@ -3764,7 +4154,7 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
url = {https://doi.org/10.1021/acs.jpca.0c07458},
urldate = {2023-04-13},
abstract = {Computations based on density functional theory (DFT) are transforming various aspects of materials research and discovery. However, the effort required to solve the central equation of DFT, namely the Kohn–Sham equation, which remains a major obstacle for studying large systems with hundreds of atoms in a practical amount of time with routine computational resources. Here, we propose a deep learning architecture that systematically learns the input–output behavior of the Kohn–Sham equation and predicts the electronic density of states, a primary output of DFT calculations, with unprecedented speed and chemical accuracy. The algorithm also adapts and progressively improves in predictive power and versatility as it is exposed to new diverse atomic configurations. We demonstrate this capability for a diverse set of carbon allotropes spanning a large configurational and phase space. The electronic density of states, along with the electronic charge density, may be used downstream to predict a variety of materials properties, bypassing the Kohn–Sham equation, leading to an ultrafast and high-fidelity DFT emulator.},
- keywords = {\_tablet,AML,carbon,defects,disordered,grid-based descriptors,materials,ML,ML-DFT,ML-ESM,NN,prediction of DOS,prediction of electron density,vacancies,VASP},
+ keywords = {AML,carbon,defects,disordered,grid-based descriptors,materials,ML,ML-DFT,ML-ESM,NN,prediction of DOS,prediction of electron density,vacancies,VASP},
file = {/Users/wasmer/Nextcloud/Zotero/del Rio et al_2020_An Efficient Deep Learning Scheme To Predict the Electronic Structure of.pdf;/Users/wasmer/Zotero/storage/46EGTQLS/acs.jpca.html}
}
@@ -3774,13 +4164,13 @@ Subject\_term\_id: computational-methods;electronic-structure;theory-and-computa
author = {Deng, Bowen and Zhong, Peichen and Jun, KyuJung and Riebesell, Janosh and Han, Kevin and Bartel, Christopher J. and Ceder, Gerbrand},
date = {2023-06-20},
eprint = {2302.14231},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2302.14231},
url = {http://arxiv.org/abs/2302.14231},
urldate = {2023-08-23},
abstract = {The simulation of large-scale systems with complex electron interactions remains one of the greatest challenges for the atomistic modeling of materials. Although classical force fields often fail to describe the coupling between electronic states and ionic rearrangements, the more accurate \textbackslash textit\{ab-initio\} molecular dynamics suffers from computational complexity that prevents long-time and large-scale simulations, which are essential to study many technologically relevant phenomena, such as reactions, ion migrations, phase transformations, and degradation. In this work, we present the Crystal Hamiltonian Graph neural Network (CHGNet) as a novel machine-learning interatomic potential (MLIP), using a graph-neural-network-based force field to model a universal potential energy surface. CHGNet is pretrained on the energies, forces, stresses, and magnetic moments from the Materials Project Trajectory Dataset, which consists of over 10 years of density functional theory static and relaxation trajectories of \$\textbackslash sim 1.5\$ million inorganic structures. The explicit inclusion of magnetic moments enables CHGNet to learn and accurately represent the orbital occupancy of electrons, enhancing its capability to describe both atomic and electronic degrees of freedom. We demonstrate several applications of CHGNet in solid-state materials, including charge-informed molecular dynamics in Li\$\_x\$MnO\$\_2\$, the finite temperature phase diagram for Li\$\_x\$FePO\$\_4\$ and Li diffusion in garnet conductors. We critically analyze the significance of including charge information for capturing appropriate chemistry, and we provide new insights into ionic systems with additional electronic degrees of freedom that can not be observed by previous MLIPs.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,ase,charge transfer,CHGNet,Database,DFT,electrochemistry,electrostatic interaction,GGA,GGA+U,ionic systems,LAMMPS,library,magnetic moment,magnetism,materials,materials project,ML,MLP,MPNN,periodic table,prediction of energy,prediction of forces,prediction of magnetic moment,prediction of stress,structure relaxation,transition metals,universal potential,VASP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Deng et al_2023_CHGNet.pdf;/Users/wasmer/Zotero/storage/IWKNTQHT/2302.html}
}
@@ -3906,7 +4296,7 @@ Subject\_term\_id: materials-science;nanoscience-and-technology},
url = {https://doi.org/10.1021/acs.chemrev.1c00022},
urldate = {2022-06-03},
abstract = {We provide an introduction to Gaussian process regression (GPR) machine-learning methods in computational materials science and chemistry. The focus of the present review is on the regression of atomistic properties: in particular, on the construction of interatomic potentials, or force fields, in the Gaussian Approximation Potential (GAP) framework; beyond this, we also discuss the fitting of arbitrary scalar, vectorial, and tensorial quantities. Methodological aspects of reference data generation, representation, and regression, as well as the question of how a data-driven model may be validated, are reviewed and critically discussed. A survey of applications to a variety of research questions in chemistry and materials science illustrates the rapid growth in the field. A vision is outlined for the development of the methodology in the years to come.},
- keywords = {\_tablet,active learning,GAP,GPR,librascal,materials,ML,ML-DFT,ML-ESM,MLP,models,molecules,prediction of electron density,prediction of LDOS,review,SA-GPR,SOAP,structure prediction,structure search,tutorial,with-code},
+ keywords = {active learning,GAP,GPR,librascal,materials,ML,ML-DFT,ML-ESM,MLP,models,molecules,prediction of electron density,prediction of LDOS,review,SA-GPR,SOAP,structure prediction,structure search,tutorial,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Deringer et al_2021_Gaussian Process Regression for Materials and Molecules.pdf;/Users/wasmer/Zotero/storage/LSTJST2A/acs.chemrev.html}
}
@@ -4045,13 +4435,13 @@ Subject\_term\_id: density-functional-theory;electronic-properties-and-materials
author = {Domina, Michelangelo and Patil, Urvesh and Cobelli, Matteo and Sanvito, Stefano},
date = {2023-06-26},
eprint = {2208.10292},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2208.10292},
url = {http://arxiv.org/abs/2208.10292},
urldate = {2023-06-27},
abstract = {We introduce a compact cluster expansion method, constructed over Jacobi and Legendre polynomials, to generate highly accurate and flexible machine-learning force fields. The constituent many-body contributions are separated, interpretable and adaptable to replicate the physical knowledge of the system. In fact, the flexibility introduced by the use of the Jacobi polynomials allows us to impose, in a natural way, constrains and symmetries to the cluster expansion. This has the effect of reducing the number of parameters needed for the fit and of enforcing desired behaviours of the potential. For instance, we show that our Jacobi-Legendre cluster expansion can be designed to generate potentials with a repulsive tail at short inter-atomic distances, without the need of imposing any external function. Our method is here continuously compared with available machine-learning potential schemes, such as the atomic cluster expansion and potentials built over the bispectrum. As an example we construct a Jacobi-Legendre potential for carbon, by training a slim and accurate model capable of describing crystalline graphite and diamond, as well as liquid and amorphous elemental carbon.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {\_tablet,ABINIT,ACE,ACE-related,AML,bispectrum,carbon,cluster expansion,DFT,forces,Jacobi-Legendre,JLP,MD,ML,ML-DFT,ML-FF,MLP,phonon,prediction of electron density},
file = {/Users/wasmer/Nextcloud/Zotero/Domina et al_2023_Cluster expansion constructed over Jacobi-Legendre polynomials for accurate.pdf;/Users/wasmer/Zotero/storage/AYU8VHC7/2208.html}
}
@@ -4079,13 +4469,13 @@ Subject\_term\_id: density-functional-theory;electronic-properties-and-materials
author = {Domina, Michelangelo and Patil, Urvesh and Cobelli, Matteo and Sanvito, Stefano},
date = {2022-08-22},
eprint = {2208.10292},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2208.10292},
url = {http://arxiv.org/abs/2208.10292},
urldate = {2022-09-05},
abstract = {Inspired by the cluster expansion method, we introduce a compact machine-learning potential constructed over Jacobi and Legendre polynomials. The constituent many-body contributions are separated, fully interpretable and adaptable to replicate the physical knowledge of the system, such as a repulsive behaviour at a small inter-atomic distance. Most importantly the potential requires a small number of features to achieve accuracy comparable to that of more numerically heavy and descriptor-rich alternatives. This is here tested for an organic molecule, a crystalline solid and an amorphous compound. Furthermore, we argue that the physical interpretability of the various terms is key to the selection and training of stable potentials.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {\_tablet,ACE,descriptors,DFT,invariance,Jacobi-Legendre,JLP,linear regression,ML,ML-ESM,MLP,prediction of total energy,SNAP},
file = {/Users/wasmer/Nextcloud/Zotero/Domina et al_2022_The Jacobi-Legendre potential.pdf;/Users/wasmer/Zotero/storage/DUUKR6TZ/2208.html}
}
@@ -4115,7 +4505,7 @@ Subject\_term\_id: density-functional-theory;electronic-properties-and-materials
author = {Domina, Michelangelo and Cobelli, Matteo and Sanvito, Stefano},
date = {2022-02-23},
eprint = {2202.13773},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
url = {http://arxiv.org/abs/2202.13773},
urldate = {2022-03-23},
@@ -4129,13 +4519,13 @@ Subject\_term\_id: density-functional-theory;electronic-properties-and-materials
author = {Dornheim, Tobias and Moldabekov, Zhandos and Cangi, Attila},
date = {2021-04-07},
eprint = {2104.02941},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2104.02941},
url = {http://arxiv.org/abs/2104.02941},
urldate = {2023-02-15},
abstract = {The electronic structure in matter under extreme conditions is a challenging complex system prevalent in astrophysical objects and highly relevant for technological applications. We show how machine-learning surrogates in terms of neural networks have a profound impact on the efficient modeling of matter under extreme conditions. We demonstrate the utility of a surrogate model that is trained on \textbackslash emph\{ab initio\} quantum Monte Carlo data for various applications in the emerging field of warm dense matter research.},
- pubstate = {preprint},
+ pubstate = {prepublished},
file = {/Users/wasmer/Nextcloud/Zotero/Dornheim et al_2021_A Machine-Learning Surrogate Model for ab initio Electronic Correlations at.pdf;/Users/wasmer/Zotero/storage/F4428MJB/2104.html}
}
@@ -4246,7 +4636,7 @@ Subject\_term\_id: density-functional-theory;electronic-properties-and-materials
url = {https://link.aps.org/doi/10.1103/PhysRevB.99.014104},
urldate = {2022-05-11},
abstract = {The atomic cluster expansion is developed as a complete descriptor of the local atomic environment, including multicomponent materials, and its relation to a number of other descriptors and potentials is discussed. The effort for evaluating the atomic cluster expansion is shown to scale linearly with the number of neighbors, irrespective of the order of the expansion. Application to small Cu clusters demonstrates smooth convergence of the atomic cluster expansion to meV accuracy. By introducing nonlinear functions of the atomic cluster expansion an interatomic potential is obtained that is comparable in accuracy to state-of-the-art machine learning potentials. Because of the efficient convergence of the atomic cluster expansion relevant subspaces can be sampled uniformly and exhaustively. This is demonstrated by testing against a large database of density functional theory calculations for copper.},
- keywords = {\_tablet,ACE,descriptors,original publication},
+ keywords = {ACE,descriptors,original publication},
file = {/Users/wasmer/Nextcloud/Zotero/Drautz_2019_Atomic cluster expansion for accurate and transferable interatomic potentials.pdf;/Users/wasmer/Zotero/storage/HNR9ZCLL/Drautz_2019_Atomic cluster expansion for accurate and transferable interatomic potentials.pdf;/Users/wasmer/Zotero/storage/NMAUF3NJ/PhysRevB.99.html}
}
@@ -4259,7 +4649,7 @@ Subject\_term\_id: density-functional-theory;electronic-properties-and-materials
volume = {102},
number = {2},
doi = {10.1103/PhysRevB.102.024104},
- keywords = {\_tablet,ACE,descriptors,magnetism,ML,spin-dependent},
+ keywords = {ACE,descriptors,magnetism,ML,spin-dependent},
file = {/Users/wasmer/Nextcloud/Zotero/Drautz_2020_Atomic cluster expansion of scalar, vectorial, and tensorial properties.pdf;/Users/wasmer/Zotero/storage/9W2WE4WX/PhysRevB.102.html}
}
@@ -4268,14 +4658,14 @@ Subject\_term\_id: density-functional-theory;electronic-properties-and-materials
author = {Drautz, Ralf and Ortner, Christoph},
date = {2022-06-22},
eprint = {2206.11375},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2206.11375},
url = {http://arxiv.org/abs/2206.11375},
urldate = {2022-06-28},
abstract = {The atomic cluster expansion (ACE) has been highly successful for the parameterisation of symmetric (invariant or equivariant) properties of many-particle systems. Here, we generalize its derivation to anti-symmetric functions. We show how numerous well-known linear representations of wave functions naturally arise within this framework and we explore how recent successful nonlinear parameterisations can be further enhanced by employing ACE methodology. From this analysis we propose a wide design space of promising wave function representations.},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,Backflow,Deep learning,ML-QM,prediction of wavefunction,representation of wavefunction,Slater-Jastrow},
+ pubstate = {prepublished},
+ keywords = {ACE,Backflow,Deep learning,ML-QM,prediction of wavefunction,representation of wavefunction,Slater-Jastrow},
file = {/Users/wasmer/Nextcloud/Zotero/Drautz_Ortner_2022_Atomic cluster expansion and wave function representations.pdf;/Users/wasmer/Zotero/storage/6PTQT7NH/2206.html}
}
@@ -4368,7 +4758,7 @@ Subject\_term\_id: density-functional-theory;electronic-properties-and-materials
abstract = {Application to the Physics of Condensed Matter},
isbn = {978-3-540-32897-1},
langid = {english},
- keywords = {\_tablet,condensed matter,group theory,irreps,learning material,mathematics,rec-by-sabastian,textbook},
+ keywords = {condensed matter,group theory,irreps,learning material,mathematics,rec-by-sabastian,textbook},
file = {/Users/wasmer/Nextcloud/Zotero/Dresselhaus et al_2007_Group Theory.pdf;/Users/wasmer/Zotero/storage/GGGVNLC4/978-3-540-32899-5.html}
}
@@ -4448,10 +4838,10 @@ Subject\_term\_id: atomistic-models;computational-methods},
@unpublished{dussonAtomicClusterExpansion2021,
title = {Atomic {{Cluster Expansion}}: {{Completeness}}, {{Efficiency}} and {{Stability}}},
shorttitle = {Atomic {{Cluster Expansion}}},
- author = {Dusson, Genevieve and Bachmayr, Markus and Csanyi, Gabor and Drautz, Ralf and Etter, Simon and family=Oord, given=Cas, prefix=van der, useprefix=true and Ortner, Christoph},
+ author = {Dusson, Genevieve and Bachmayr, Markus and Csányi, Gábor and Drautz, Ralf and Etter, Simon and family=Oord, given=Cas, prefix=van der, useprefix=true and Ortner, Christoph},
date = {2021-05-12},
eprint = {1911.03550},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math},
url = {http://arxiv.org/abs/1911.03550},
urldate = {2022-05-11},
@@ -4484,13 +4874,13 @@ Subject\_term\_id: atomistic-models;computational-methods},
author = {Duval, Alexandre and Schmidt, Victor and Garcia, Alex Hernandez and Miret, Santiago and Malliaros, Fragkiskos D. and Bengio, Yoshua and Rolnick, David},
date = {2023-04-28},
eprint = {2305.05577},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2305.05577},
url = {http://arxiv.org/abs/2305.05577},
urldate = {2023-12-04},
abstract = {Applications of machine learning techniques for materials modeling typically involve functions known to be equivariant or invariant to specific symmetries. While graph neural networks (GNNs) have proven successful in such tasks, they enforce symmetries via the model architecture, which often reduces their expressivity, scalability and comprehensibility. In this paper, we introduce (1) a flexible framework relying on stochastic frame-averaging (SFA) to make any model E(3)-equivariant or invariant through data transformations. (2) FAENet: a simple, fast and expressive GNN, optimized for SFA, that processes geometric information without any symmetrypreserving design constraints. We prove the validity of our method theoretically and empirically demonstrate its superior accuracy and computational scalability in materials modeling on the OC20 dataset (S2EF, IS2RE) as well as common molecular modeling tasks (QM9, QM7-X). A package implementation is available at https://faenet.readthedocs.io.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,alternative approaches,alternative for equivariance,AML,approximative equivariance,equivariant,frame averaging,GNN,library,materials,ML,todo-tagging,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Duval et al_2023_FAENet.pdf;/Users/wasmer/Zotero/storage/I9RAQSSH/2305.html}
}
@@ -4500,23 +4890,40 @@ Subject\_term\_id: atomistic-models;computational-methods},
author = {Duval, Alexandre and Mathis, Simon V. and Joshi, Chaitanya K. and Schmidt, Victor and Miret, Santiago and Malliaros, Fragkiskos D. and Cohen, Taco and Lio, Pietro and Bengio, Yoshua and Bronstein, Michael},
date = {2023-12-12},
eprint = {2312.07511},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, q-bio, stat},
doi = {10.48550/arXiv.2312.07511},
url = {http://arxiv.org/abs/2312.07511},
urldate = {2024-01-13},
abstract = {Recent advances in computational modelling of atomic systems, spanning molecules, proteins, and materials, represent them as geometric graphs with atoms embedded as nodes in 3D Euclidean space. In these graphs, the geometric attributes transform according to the inherent physical symmetries of 3D atomic systems, including rotations and translations in Euclidean space, as well as node permutations. In recent years, Geometric Graph Neural Networks have emerged as the preferred machine learning architecture powering applications ranging from protein structure prediction to molecular simulations and material generation. Their specificity lies in the inductive biases they leverage -- such as physical symmetries and chemical properties -- to learn informative representations of these geometric graphs. In this opinionated paper, we provide a comprehensive and self-contained overview of the field of Geometric GNNs for 3D atomic systems. We cover fundamental background material and introduce a pedagogical taxonomy of Geometric GNN architectures:(1) invariant networks, (2) equivariant networks in Cartesian basis, (3) equivariant networks in spherical basis, and (4) unconstrained networks. Additionally, we outline key datasets and application areas and suggest future research directions. The objective of this work is to present a structured perspective on the field, making it accessible to newcomers and aiding practitioners in gaining an intuition for its mathematical abstractions.},
- pubstate = {preprint},
- keywords = {\_tablet,Allegro,AML,benchmarking,best practices,biomolecules,CGCNN,DimeNet,E(n),e3nn,EGNN,equivariant,equivariant alternative,GemNet,geometric deep learning,geometric GNNs,GNN,graph ML,invariance,MACE,materials,MEGNet,ML,ML theory,model comparison,molecules,MPNN,NequIP,PAiNN,reference,review,SchNet,SE(3),SO(3),SphereNet,symmetry,tensor field,TensorNet,theory,transformer,tutorial,with-data},
+ pubstate = {prepublished},
+ keywords = {Allegro,AML,benchmarking,best practices,biomolecules,CGCNN,DimeNet,E(n),e3nn,EGNN,equivariant,equivariant alternative,GemNet,geometric deep learning,geometric GNNs,GNN,graph ML,invariance,MACE,materials,MEGNet,ML,ML theory,model comparison,molecules,MPNN,NequIP,PAiNN,reference,review,review-of-AML,SchNet,SE(3),SO(3),SphereNet,symmetry,tensor field,TensorNet,theory,transformer,tutorial,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Duval et al_2023_A Hitchhiker's Guide to Geometric GNNs for 3D Atomic Systems.pdf;/Users/wasmer/Zotero/storage/2CWI5WGP/2312.html}
}
+@online{duvalHitchhikerGuideGeometric2024,
+ title = {A {{Hitchhiker}}'s {{Guide}} to {{Geometric GNNs}} for {{3D Atomic Systems}}},
+ author = {Duval, Alexandre and Mathis, Simon V. and Joshi, Chaitanya K. and Schmidt, Victor and Miret, Santiago and Malliaros, Fragkiskos D. and Cohen, Taco and Liò, Pietro and Bengio, Yoshua and Bronstein, Michael},
+ date = {2024-03-13},
+ eprint = {2312.07511},
+ eprinttype = {arXiv},
+ eprintclass = {cs, q-bio, stat},
+ doi = {10.48550/arXiv.2312.07511},
+ url = {http://arxiv.org/abs/2312.07511},
+ urldate = {2024-05-29},
+ abstract = {Recent advances in computational modelling of atomic systems, spanning molecules, proteins, and materials, represent them as geometric graphs with atoms embedded as nodes in 3D Euclidean space. In these graphs, the geometric attributes transform according to the inherent physical symmetries of 3D atomic systems, including rotations and translations in Euclidean space, as well as node permutations. In recent years, Geometric Graph Neural Networks have emerged as the preferred machine learning architecture powering applications ranging from protein structure prediction to molecular simulations and material generation. Their specificity lies in the inductive biases they leverage - such as physical symmetries and chemical properties - to learn informative representations of these geometric graphs. In this opinionated paper, we provide a comprehensive and self-contained overview of the field of Geometric GNNs for 3D atomic systems. We cover fundamental background material and introduce a pedagogical taxonomy of Geometric GNN architectures: (1) invariant networks, (2) equivariant networks in Cartesian basis, (3) equivariant networks in spherical basis, and (4) unconstrained networks. Additionally, we outline key datasets and application areas and suggest future research directions. The objective of this work is to present a structured perspective on the field, making it accessible to newcomers and aiding practitioners in gaining an intuition for its mathematical abstractions.},
+ pubstate = {prepublished},
+ version = {2},
+ keywords = {Allegro,AML,benchmarking,best practices,biomolecules,CGCNN,DimeNet,E(n),e3nn,EGNN,equivariant,equivariant alternative,GemNet,geometric deep learning,geometric GNNs,GNN,graph ML,invariance,MACE,materials,MEGNet,ML,ML theory,model comparison,molecules,MPNN,NequIP,PAiNN,reference,reivew-of-AML,review,SchNet,SE(3),SO(3),SphereNet,Statistics - Machine Learning,symmetry,tensor field,TensorNet,theory,transformer,tutorial,with-data},
+ file = {/Users/wasmer/Nextcloud/Zotero/Duval et al_2024_A Hitchhiker's Guide to Geometric GNNs for 3D Atomic Systems.pdf;/Users/wasmer/Zotero/storage/ITAP97UZ/2312.html}
+}
+
@unpublished{dymLowDimensionalInvariant2022,
title = {Low {{Dimensional Invariant Embeddings}} for {{Universal Geometric Learning}}},
author = {Dym, Nadav and Gortler, Steven J.},
date = {2022-05-05},
eprint = {2205.02956},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math},
publisher = {arXiv},
doi = {10.48550/arXiv.2205.02956},
@@ -4543,7 +4950,7 @@ Subject\_term\_id: atomistic-models;computational-methods},
urldate = {2021-10-04},
abstract = {The modern version of the KKR (Korringa–Kohn–Rostoker) method represents the electronic structure of a system directly and efficiently in terms of its single-particle Green's function (GF). This is in contrast to its original version and many other traditional wave-function-based all-electron band structure methods dealing with periodically ordered solids. Direct access to the GF results in several appealing features. In addition, a wide applicability of the method is achieved by employing multiple scattering theory. The basic ideas behind the resulting KKR-GF method are outlined and the different techniques to deal with the underlying multiple scattering problem are reviewed. Furthermore, various applications of the KKR-GF method are reviewed in some detail to demonstrate the remarkable flexibility of the approach. Special attention is devoted to the numerous developments of the KKR-GF method, that have been contributed in recent years by a number of work groups, in particular in the following fields: embedding schemes for atoms, clusters and surfaces, magnetic response functions and anisotropy, electronic and spin-dependent transport, dynamical mean field theory, various kinds of spectroscopies, as well as first-principles determination of model parameters.},
langid = {english},
- keywords = {\_tablet,KKR,review},
+ keywords = {KKR,review},
file = {/Users/wasmer/Nextcloud/Zotero/Ebert et al_2011_Calculating condensed matter properties using the KKR-Green's function.pdf}
}
@@ -4552,7 +4959,7 @@ Subject\_term\_id: atomistic-models;computational-methods},
author = {Eckhoff, Marco and Behler, Jörg},
date = {2021-04-29},
eprint = {2104.14439},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
url = {http://arxiv.org/abs/2104.14439},
urldate = {2021-05-18},
@@ -4597,13 +5004,13 @@ Subject\_term\_id: atomistic-models;computational-methods},
author = {Eisenbach, Markus and Karabin, Mariia and Pasini, Massimiliano Lupo and Yin, Junqi},
date = {2022-07-29},
eprint = {2207.10144},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2207.10144},
url = {http://arxiv.org/abs/2207.10144},
urldate = {2023-09-19},
abstract = {The investigation of finite temperature properties using Monte-Carlo (MC) methods requires a large number of evaluations of the system's Hamiltonian to sample the phase space needed to obtain physical observables as function of temperature. DFT calculations can provide accurate evaluations of the energies, but they are too computationally expensive for routine simulations. To circumvent this problem, machine-learning (ML) based surrogate models have been developed and implemented on high-performance computing (HPC) architectures. In this paper, we describe two ML methods (linear mixing model and HydraGNN) as surrogates for first principles density functional theory (DFT) calculations with classical MC simulations. These two surrogate models are used to learn the dependence of target physical properties from complex compositions and interactions of their constituents. We present the predictive performance of these two surrogate models with respect to their complexity while avoiding the danger of overfitting the model. An important aspect of our approach is the periodic retraining with newly generated first principles data based on the progressive exploration of the system's phase space by the MC simulation. The numerical results show that HydraGNN model attains superior predictive performance compared to the linear mixing model for magnetic alloy materials.},
- pubstate = {preprint},
+ pubstate = {prepublished},
version = {2},
keywords = {active learning,alloys,AML,binary systems,DFT,Ferromagnetism,GNN,HEA,HydraGNN,linear mixing model,magnetism,MC,ML,ML-DFT,multi-task learning,Multiple scattering theory,prediction of charge transfer,prediction of energy,prediction of magnetic moment,surrogate model},
file = {/Users/wasmer/Nextcloud/Zotero/Eisenbach et al_2022_Machine Learning for First Principles Calculations of Material Properties for.pdf;/Users/wasmer/Zotero/storage/ZHJES8AM/2207.html}
@@ -4675,6 +5082,22 @@ Junqi Yin\\
keywords = {/unread,todo-tagging}
}
+@online{elijosiusZeroShotMolecular2024,
+ title = {Zero {{Shot Molecular Generation}} via {{Similarity Kernels}}},
+ author = {Elijošius, Rokas and Zills, Fabian and Batatia, Ilyes and Norwood, Sam Walton and Kovács, Dávid Péter and Holm, Christian and Csányi, Gábor},
+ date = {2024-02-13},
+ eprint = {2402.08708},
+ eprinttype = {arXiv},
+ eprintclass = {physics},
+ doi = {10.48550/arXiv.2402.08708},
+ url = {http://arxiv.org/abs/2402.08708},
+ urldate = {2024-06-14},
+ abstract = {Generative modelling aims to accelerate the discovery of novel chemicals by directly proposing structures with desirable properties. Recently, score-based, or diffusion, generative models have significantly outperformed previous approaches. Key to their success is the close relationship between the score and physical force, allowing the use of powerful equivariant neural networks. However, the behaviour of the learnt score is not yet well understood. Here, we analyse the score by training an energy-based diffusion model for molecular generation. We find that during the generation the score resembles a restorative potential initially and a quantum-mechanical force at the end. In between the two endpoints, it exhibits special properties that enable the building of large molecules. Using insights from the trained model, we present Similarity-based Molecular Generation (SiMGen), a new method for zero shot molecular generation. SiMGen combines a time-dependent similarity kernel with descriptors from a pretrained machine learning force field to generate molecules without any further training. Our approach allows full control over the molecular shape through point cloud priors and supports conditional generation. We also release an interactive web tool that allows users to generate structures with SiMGen online (https://zndraw.icp.uni-stuttgart.de).},
+ pubstate = {prepublished},
+ keywords = {/unread,AML,generative models,kernel methods,library,MACE,MD,ML,MLP,molecules,organic chemistry,pretrained models,similarity kernel,similarity measure,visualization,with-code,with-demo,zero-shot learning},
+ file = {/Users/wasmer/Nextcloud/Zotero/Elijošius et al_2024_Zero Shot Molecular Generation via Similarity Kernels.pdf;/Users/wasmer/Zotero/storage/L7ZJNP6X/2402.html}
+}
+
@article{ellisAcceleratingFinitetemperatureKohnSham2021,
title = {Accelerating Finite-Temperature {{Kohn-Sham}} Density Functional Theory with Deep Neural Networks},
author = {Ellis, J. A. and Fiedler, L. and Popoola, G. A. and Modine, N. A. and Stephens, J. A. and Thompson, A. P. and Cangi, A. and Rajamanickam, S.},
@@ -4707,6 +5130,23 @@ Junqi Yin\\
file = {/Users/wasmer/Nextcloud/Zotero/Erwin_Edkins_2023_White Paper - Leveraging Physics-Based Models and AI for new Material.pdf;/Users/wasmer/Zotero/storage/QJBC8LD2/white-paper-leveraging-physics-based-models-and-ai-for-new-material-development.html}
}
+@article{evansDevelopmentsApplicationsOPTIMADE2024,
+ title = {Developments and Applications of the {{OPTIMADE API}} for Materials Discovery, Design, and Data Exchange},
+ author = {Evans, Matthew L. and Bergsma, Johan and Merkys, Andrius and Andersen, Casper W. and Andersson, Oskar B. and Beltrán, Daniel and Blokhin, Evgeny and Boland, Tara M. and Balderas, Rubén Castañeda and Choudhary, Kamal and DÃaz, Alberto DÃaz and GarcÃa, Rodrigo DomÃnguez and Eckert, Hagen and Eimre, Kristjan and Montero, MarÃa Elena Fuentes and Krajewski, Adam M. and Mortensen, Jens Jørgen and Duarte, José Manuel Nápoles and Pietryga, Jacob and Qi, Ji and Carrillo, Felipe de Jesús Trejo and Vaitkus, Antanas and Yu, Jusong and Zettel, Adam and family=Castro, given=Pedro Baptista, prefix=de, useprefix=false and Carlsson, Johan and Cerqueira, Tiago F. T. and Divilov, Simon and Hajiyani, Hamidreza and Hanke, Felix and Jose, Kevin and Oses, Corey and Riebesell, Janosh and Schmidt, Jonathan and Winston, Donald and Xie, Christen and Yang, Xiaoyu and Bonella, Sara and Botti, Silvana and Curtarolo, Stefano and Draxl, Claudia and Cobas, Luis Edmundo Fuentes and Hospital, Adam and Liu, Zi-Kui and Marques, Miguel A. L. and Marzari, Nicola and Morris, Andrew J. and Ong, Shyue Ping and Orozco, Modesto and Persson, Kristin A. and Thygesen, Kristian S. and Wolverton, Chris and Scheidgen, Markus and Toher, Cormac and Conduit, Gareth J. and Pizzi, Giovanni and Gražulis, Saulius and Rignanese, Gian-Marco and Armiento, Rickard},
+ date = {2024-04-18},
+ journaltitle = {Digital Discovery},
+ shortjournal = {Digital Discovery},
+ publisher = {RSC},
+ issn = {2635-098X},
+ doi = {10.1039/D4DD00039K},
+ url = {https://pubs.rsc.org/en/content/articlelanding/2024/dd/d4dd00039k},
+ urldate = {2024-06-26},
+ abstract = {The Open Databases Integration for Materials Design (OPTIMADE) application programming interface (API) empowers users with holistic access to a growing federation of databases, enhancing the accessibility and discoverability of materials and chemical data. Since the first release of the OPTIMADE specification (v1.0), the API has undergone significant development, leading to the v1.2 release, and has underpinned multiple scientific studies. In this work, we highlight the latest features of the API format, accompanying software tools, and provide an update on the implementation of OPTIMADE in contributing materials databases. We end by providing several use cases that demonstrate the utility of the OPTIMADE API in materials research that continue to drive its ongoing development.},
+ langid = {english},
+ keywords = {/unread,AFLOW,Alexandria database,CMR database,Crystallography Open Database,interoperability,JARVIS,library,materials,Materials Cloud,materials database,materials informatics,materials project,matterverse,MPDS database,NOMAD,OPTIMADE,OQMD,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Evans et al_2024_Developments and applications of the OPTIMADE API for materials discovery,.pdf;/Users/wasmer/Zotero/storage/ZW48R3TC/Evans et al. - 2024 - Developments and applications of the OPTIMADE API .pdf}
+}
+
@article{evansGroupTheory2004,
title = {Group {{Theory}}},
author = {Evans, Tim S. and Vvedensky, Dimitri D.},
@@ -4851,7 +5291,6 @@ Junqi Yin\\
urldate = {2021-12-02},
isbn = {978-0-7503-1490-9},
langid = {english},
- keywords = {\_tablet},
file = {/Users/wasmer/Nextcloud/Zotero/Multiple Scattering Theory.pdf;/Users/wasmer/Zotero/storage/UYLUXULV/978-0-7503-1490-9.html}
}
@@ -4888,13 +5327,13 @@ Junqi Yin\\
author = {Feng, Rui and Zhu, Qi and Tran, Huan and Chen, Binghong and Toland, Aubrey and Ramprasad, Rampi and Zhang, Chao},
date = {2023-08-23},
eprint = {2308.14759},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, q-bio},
doi = {10.48550/arXiv.2308.14759},
url = {http://arxiv.org/abs/2308.14759},
urldate = {2023-09-22},
abstract = {Recent works have shown the promise of learning pre-trained models for 3D molecular representation. However, existing pre-training models focus predominantly on equilibrium data and largely overlook off-equilibrium conformations. It is challenging to extend these methods to off-equilibrium data because their training objective relies on assumptions of conformations being the local energy minima. We address this gap by proposing a force-centric pretraining model for 3D molecular conformations covering both equilibrium and off-equilibrium data. For off-equilibrium data, our model learns directly from their atomic forces. For equilibrium data, we introduce zero-force regularization and forced-based denoising techniques to approximate near-equilibrium forces. We obtain a unified pre-trained model for 3D molecular representation with over 15 million diverse conformations. Experiments show that, with our pre-training objective, we increase forces accuracy by around 3 times compared to the un-pre-trained Equivariant Transformer model. By incorporating regularizations on equilibrium data, we solved the problem of unstable MD simulations in vanilla Equivariant Transformers, achieving state-of-the-art simulation performance with 2.45 times faster inference time than NequIP. As a powerful molecular encoder, our pre-trained model achieves on-par performance with state-of-the-art property prediction tasks.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,ANI1-x,fine-tuning,finetuning,MD,MD17,ML,MLP,molecules,NequIP,out-of-equilibrium,pretrained models,QM9,representation learning},
file = {/Users/wasmer/Zotero/storage/IAZ2DY2B/Feng et al. - 2023 - May the Force be with You Unified Force-Centric P.pdf;/Users/wasmer/Zotero/storage/UG7IJTUA/2308.html}
}
@@ -4905,13 +5344,13 @@ Junqi Yin\\
author = {Feng, Rui and Tran, Huan and Toland, Aubrey and Chen, Binghong and Zhu, Qi and Ramprasad, Rampi and Zhang, Chao},
date = {2023-09-01},
eprint = {2309.00585},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2309.00585},
url = {http://arxiv.org/abs/2309.00585},
urldate = {2023-09-22},
abstract = {Polymer simulation with both accuracy and efficiency is a challenging task. Machine learning (ML) forcefields have been developed to achieve both the accuracy of ab initio methods and the efficiency of empirical force fields. However, existing ML force fields are usually limited to single-molecule settings, and their simulations are not robust enough. In this paper, we present PolyGET, a new framework for Polymer Forcefields with Generalizable Equivariant Transformers. PolyGET is designed to capture complex quantum interactions between atoms and generalize across various polymer families, using a deep learning model called Equivariant Transformers. We propose a new training paradigm that focuses exclusively on optimizing forces, which is different from existing methods that jointly optimize forces and energy. This simple force-centric objective function avoids competing objectives between energy and forces, thereby allowing for learning a unified forcefield ML model over different polymer families. We evaluated PolyGET on a large-scale dataset of 24 distinct polymer types and demonstrated state-of-the-art performance in force accuracy and robust MD simulations. Furthermore, PolyGET can simulate large polymers with high fidelity to the reference ab initio DFT method while being able to generalize to unseen polymers.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,benchmarking,EGNN,equivariant,MD,ML,ML-FF,MLP,NequIP,PolyGET,polymers,SchNet,TorchMDNet,transformer},
file = {/Users/wasmer/Zotero/storage/KZMGIR4P/Feng et al. - 2023 - PolyGET Accelerating Polymer Simulations by Accur.pdf;/Users/wasmer/Zotero/storage/AYKT7AGJ/2309.html}
}
@@ -4950,13 +5389,13 @@ Junqi Yin\\
author = {Fiedler, Lenz and Moldabekov, Zhandos A. and Shao, Xuecheng and Jiang, Kaili and Dornheim, Tobias and Pavanello, Michele and Cangi, Attila},
date = {2022-09-02},
eprint = {2206.03754},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2206.03754},
url = {http://arxiv.org/abs/2206.03754},
urldate = {2023-02-15},
abstract = {We introduce a practical hybrid approach that combines orbital-free density functional theory (DFT) with Kohn-Sham DFT for speeding up first-principles molecular dynamics simulations. Equilibrated ionic configurations are generated using orbital-free DFT for subsequent Kohn-Sham DFT molecular dynamics. This leads to a massive reduction of the simulation time without any sacrifice in accuracy. We assess this finding across systems of different sizes and temperature, up to the warm dense matter regime. To that end, we use the cosine distance between the time series of radial distribution functions representing the ionic configurations. Likewise, we show that the equilibrated ionic configurations from this hybrid approach significantly enhance the accuracy of machine-learning models that replace Kohn-Sham DFT. Our hybrid scheme enables systematic first-principles simulations of warm dense matter that are otherwise hampered by the large numbers of atoms and the prevalent high temperatures. Moreover, our finding provides an additional motivation for developing kinetic and noninteracting free energy functionals for orbital-free DFT.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AIMD,AML,DFT,finite-temperature DFT,LDA,library,MALA,MD,ML,ML-DFT,ML-ESM,OF-DFT,Quantum ESPRESSO,single-element,VASP,warm dense matter,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Fiedler et al_2022_Accelerating Equilibration in First-Principles Molecular Dynamics with.pdf;/Users/wasmer/Zotero/storage/TA7XVJUP/2206.html}
}
@@ -4966,12 +5405,12 @@ Junqi Yin\\
author = {Fiedler, Lenz and Shah, Karan and Bussmann, Michael and Cangi, Attila},
date = {2021-10-03},
eprint = {2110.00997},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
url = {http://arxiv.org/abs/2110.00997},
urldate = {2021-11-17},
abstract = {With the growth of computational resources, the scope of electronic structure simulations has increased greatly. Artificial intelligence and robust data analysis hold the promise to accelerate large-scale simulations and their analysis to hitherto unattainable scales. Machine learning is a rapidly growing field for the processing of such complex datasets. It has recently gained traction in the domain of electronic structure simulations, where density functional theory takes the prominent role of the most widely used electronic structure method. Thus, DFT calculations represent one of the largest loads on academic high-performance computing systems across the world. Accelerating these with machine learning can reduce the resources required and enables simulations of larger systems. Hence, the combination of density functional theory and machine learning has the potential to rapidly advance electronic structure applications such as in-silico materials discovery and the search for new chemical reaction pathways. We provide the theoretical background of both density functional theory and machine learning on a generally accessible level. This serves as the basis of our comprehensive review including research articles up to December 2020 in chemistry and materials science that employ machine-learning techniques. In our analysis, we categorize the body of research into main threads and extract impactful results. We conclude our review with an outlook on exciting research directions in terms of a citation analysis.},
- keywords = {\_tablet,citation analysis,DFT,literature analysis,ML,ML-DFT,ML-ESM,review},
+ keywords = {citation analysis,DFT,literature analysis,ML,ML-DFT,ML-ESM,review},
file = {/Users/wasmer/Nextcloud/Zotero/Fiedler et al_2021_A Deep Dive into Machine Learning Density Functional Theory for Materials.pdf;/Users/wasmer/Zotero/storage/2XW6IGEA/2110.html}
}
@@ -4980,13 +5419,13 @@ Junqi Yin\\
author = {Fiedler, Lenz and Shah, Karan and Bussmann, Michael and Cangi, Attila},
date = {2022-02-25},
eprint = {2110.00997},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2110.00997},
url = {http://arxiv.org/abs/2110.00997},
urldate = {2023-02-15},
abstract = {With the growth of computational resources, the scope of electronic structure simulations has increased greatly. Artificial intelligence and robust data analysis hold the promise to accelerate large-scale simulations and their analysis to hitherto unattainable scales. Machine learning is a rapidly growing field for the processing of such complex datasets. It has recently gained traction in the domain of electronic structure simulations, where density functional theory takes the prominent role of the most widely used electronic structure method. Thus, DFT calculations represent one of the largest loads on academic high-performance computing systems across the world. Accelerating these with machine learning can reduce the resources required and enables simulations of larger systems. Hence, the combination of density functional theory and machine learning has the potential to rapidly advance electronic structure applications such as in-silico materials discovery and the search for new chemical reaction pathways. We provide the theoretical background of both density functional theory and machine learning on a generally accessible level. This serves as the basis of our comprehensive review including research articles up to December 2020 in chemistry and materials science that employ machine-learning techniques. In our analysis, we categorize the body of research into main threads and extract impactful results. We conclude our review with an outlook on exciting research directions in terms of a citation analysis.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {citation analysis,DFT,literature analysis,ML,ML-DFA,ML-DFT,ML-ESM,prediction of electron density,review},
file = {/Users/wasmer/Nextcloud/Zotero/Fiedler et al_2022_A Deep Dive into Machine Learning Density Functional Theory for Materials.pdf;/Users/wasmer/Zotero/storage/NUWIJFTB/2110.html}
}
@@ -5014,13 +5453,13 @@ Junqi Yin\\
author = {Fiedler, Lenz and Modine, Normand A. and Miller, Kyle D. and Cangi, Attila},
date = {2023-06-09},
eprint = {2306.06032},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2306.06032},
url = {http://arxiv.org/abs/2306.06032},
urldate = {2023-06-13},
abstract = {We introduce machine learning (ML) models that predict the electronic structure of materials across a wide temperature range. Our models employ neural networks and are trained on density functional theory (DFT) data. Unlike other ML models that use DFT data, our models directly predict the local density of states (LDOS) of the electronic structure. This provides several advantages, including access to multiple observables such as the electronic density and electronic total free energy. Moreover, our models account for both the electronic and ionic temperatures independently, making them ideal for applications like laser-heating of matter. We validate the efficacy of our LDOS-based models on a metallic test system. They accurately capture energetic effects induced by variations in ionic and electronic temperatures over a broad temperature range, even when trained on a subset of these temperatures. These findings open up exciting opportunities for investigating the electronic structure of materials under both ambient and extreme conditions.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {aluminium,AML,descriptors,DFT,finite-temperature DFT,library,MALA,Metals and alloys,ML,ML-DFT,ML-ESM,model reporting,PAW,PBE,prediction of electron density,prediction of LDOS,prediction of total energy,Quantum ESPRESSO,transfer learning,VASP,warm dense matter,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Fiedler et al_2023_Machine learning the electronic structure of matter across temperatures.pdf;/Users/wasmer/Zotero/storage/YX8Z9U2B/2306.html}
}
@@ -5050,13 +5489,13 @@ Junqi Yin\\
author = {Fiedler, Lenz and Modine, Normand A. and Schmerler, Steve and Vogel, Dayton J. and Popoola, Gabriel A. and Thompson, Aidan P. and Rajamanickam, Sivasankaran and Cangi, Attila},
date = {2022-12-08},
eprint = {2210.11343},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2210.11343},
url = {http://arxiv.org/abs/2210.11343},
urldate = {2023-02-15},
abstract = {The properties of electrons in matter are of fundamental importance. They give rise to virtually all molecular and material properties and determine the physics at play in objects ranging from semiconductor devices to the interior of giant gas planets. Modeling and simulation of such diverse applications rely primarily on density functional theory (DFT), which has become the principal method for predicting the electronic structure of matter. While DFT calculations have proven to be very useful to the point of being recognized with a Nobel prize in 1998, their computational scaling limits them to small systems. We have developed a machine learning framework for predicting the electronic structure on any length scale. It shows up to three orders of magnitude speedup on systems where DFT is tractable and, more importantly, enables predictions on scales where DFT calculations are infeasible. Our work demonstrates how machine learning circumvents a long-standing computational bottleneck and advances science to frontiers intractable with any current solutions. This unprecedented modeling capability opens up an inexhaustible range of applications in astrophysics, novel materials discovery, and energy solutions for a sustainable future.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AIMD,AML,bispectrum,defect 2D,defects,descriptors,DFT,disordered,finite-temperature DFT,grid-based descriptors,LAMMPS,library,linear scaling,linear-scaling DFT,MALA,ML,ML-DFT,ML-ESM,prediction of electron density,prediction of energy,prediction of LDOS,Quantum ESPRESSO,scaling,VASP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Fiedler et al_2022_Predicting electronic structures at any length scale with machine learning.pdf;/Users/wasmer/Zotero/storage/9AYDDQ8T/2210.html}
}
@@ -5091,7 +5530,7 @@ Junqi Yin\\
volume = {3},
number = {4},
eprint = {2202.09186},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
pages = {045008},
issn = {2632-2153},
@@ -5099,7 +5538,7 @@ Junqi Yin\\
url = {http://arxiv.org/abs/2202.09186},
urldate = {2023-02-15},
abstract = {A myriad of phenomena in materials science and chemistry rely on quantum-level simulations of the electronic structure in matter. While moving to larger length and time scales has been a pressing issue for decades, such large-scale electronic structure calculations are still challenging despite modern software approaches and advances in high-performance computing. The silver lining in this regard is the use of machine learning to accelerate electronic structure calculations -- this line of research has recently gained growing attention. The grand challenge therein is finding a suitable machine-learning model during a process called hyperparameter optimization. This, however, causes a massive computational overhead in addition to that of data generation. We accelerate the construction of neural network models by roughly two orders of magnitude by circumventing excessive training during the hyperparameter optimization phase. We demonstrate our workflow for Kohn-Sham density functional theory, the most popular computational method in materials science and chemistry.},
- keywords = {\_tablet,AML,Deep learning,hyperparameters,hyperparameters optimization,library,MALA,ML,ML-DFT,ML-ESM,Optuna,prediction of electron density,prediction of LDOS,with-code},
+ keywords = {AML,Deep learning,hyperparameters,hyperparameters optimization,library,MALA,ML,ML-DFT,ML-ESM,Optuna,prediction of electron density,prediction of LDOS,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Fiedler et al_2022_Training-free hyperparameter optimization of neural networks for electronic.pdf;/Users/wasmer/Zotero/storage/6BKNM2VX/2202.html}
}
@@ -5123,22 +5562,76 @@ Junqi Yin\\
file = {/Users/wasmer/Nextcloud/Zotero/Fiedler et al_2022_Training-free hyperparameter optimization of neural networks for electronic2.pdf}
}
+@article{finkbeinerGeneratingMinimalTraining2024,
+ title = {Generating {{Minimal Training Sets}} for {{Machine Learned Potentials}}},
+ author = {Finkbeiner, Jan and Tovey, Samuel and Holm, Christian},
+ date = {2024-04-15},
+ journaltitle = {Physical Review Letters},
+ shortjournal = {Phys. Rev. Lett.},
+ volume = {132},
+ number = {16},
+ pages = {167301},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevLett.132.167301},
+ url = {https://link.aps.org/doi/10.1103/PhysRevLett.132.167301},
+ urldate = {2024-06-14},
+ abstract = {This Letter presents a novel approach for identifying uncorrelated atomic configurations from extensive datasets with a nonstandard neural network workflow known as random network distillation (RND) for training machine-learned interatomic potentials (MLPs). This method is coupled with a DFT workflow wherein initial data are generated with cheaper classical methods before only the minimal subset is passed to a more computationally expensive ab initio calculation. This benefits training not only by reducing the number of expensive DFT calculations required but also by providing a pathway to the use of more accurate quantum mechanical calculations. The method’s efficacy is demonstrated by constructing machine-learned interatomic potentials for the molten salts KCl and NaCl. Our RND method allows accurate models to be fit on minimal datasets, as small as 32 configurations, reducing the required structures by at least 1 order of magnitude compared to alternative methods. This reduction in dataset sizes not only substantially reduces computational overhead for training data generation but also provides a more comprehensive starting point for active-learning procedures.},
+ keywords = {/unread,AML,database generation,ML,MLP,small data},
+ file = {/Users/wasmer/Nextcloud/Zotero/Finkbeiner et al_2024_Generating Minimal Training Sets for Machine Learned Potentials.pdf;/Users/wasmer/Zotero/storage/FZM7UD4D/PhysRevLett.132.html}
+}
+
+@article{finocchioRoadmapUnconventionalComputing2024,
+ title = {Roadmap for Unconventional Computing with Nanotechnology},
+ author = {Finocchio, Giovanni and Incorvia, Jean Anne C. and Friedman, Joseph S. and Yang, Qu and Giordano, Anna and Grollier, Julie and Yang, Hyunsoo and Ciubotaru, Florin and Chumak, Andrii V. and Naeemi, Azad J. and Cotofana, Sorin D. and Tomasello, Riccardo and Panagopoulos, Christos and Carpentieri, Mario and Lin, Peng and Pan, Gang and Yang, J. Joshua and Todri-Sanial, Aida and Boschetto, Gabriele and Makasheva, Kremena and Sangwan, Vinod K. and Trivedi, Amit Ranjan and Hersam, Mark C. and Camsari, Kerem Y. and McMahon, Peter L. and Datta, Supriyo and Koiller, Belita and Aguilar, Gabriel H. and Temporão, Guilherme P. and Rodrigues, Davi R. and Sunada, Satoshi and Everschor-Sitte, Karin and Tatsumura, Kosuke and Goto, Hayato and Puliafito, Vito and Åkerman, Johan and Takesue, Hiroki and Ventra, Massimiliano Di and Pershin, Yuriy V. and Mukhopadhyay, Saibal and Roy, Kaushik and Wang, I.-Ting and Kang, Wang and Zhu, Yao and Kaushik, Brajesh Kumar and Hasler, Jennifer and Ganguly, Samiran and Ghosh, Avik W. and Levy, William and Roychowdhury, Vwani and Bandyopadhyay, Supriyo},
+ date = {2024-03},
+ journaltitle = {Nano Futures},
+ shortjournal = {Nano Futures},
+ volume = {8},
+ number = {1},
+ pages = {012001},
+ publisher = {IOP Publishing},
+ issn = {2399-1984},
+ doi = {10.1088/2399-1984/ad299a},
+ url = {https://dx.doi.org/10.1088/2399-1984/ad299a},
+ urldate = {2024-07-17},
+ abstract = {In the ‘Beyond Moore’s Law’ era, with increasing edge intelligence, domain-specific computing embracing unconventional approaches will become increasingly prevalent. At the same time, adopting a variety of nanotechnologies will offer benefits in energy cost, computational speed, reduced footprint, cyber resilience, and processing power. The time is ripe for a roadmap for unconventional computing with nanotechnologies to guide future research, and this collection aims to fill that need. The authors provide a comprehensive roadmap for neuromorphic computing using electron spins, memristive devices, two-dimensional nanomaterials, nanomagnets, and various dynamical systems. They also address other paradigms such as Ising machines, Bayesian inference engines, probabilistic computing with p-bits, processing in memory, quantum memories and algorithms, computing with skyrmions and spin waves, and brain-inspired computing for incremental learning and problem-solving in severely resource-constrained environments. These approaches have advantages over traditional Boolean computing based on von Neumann architecture. As the computational requirements for artificial intelligence grow 50 times faster than Moore’s Law for electronics, more unconventional approaches to computing and signal processing will appear on the horizon, and this roadmap will help identify future needs and challenges. In a very fertile field, experts in the field aim to present some of the dominant and most promising technologies for unconventional computing that will be around for some time to come. Within a holistic approach, the goal is to provide pathways for solidifying the field and guiding future impactful discoveries.},
+ langid = {english},
+ keywords = {/unread,nanomaterials,Neuromorphic,physics,quantum computing,rec-by-sanvito,review,review-of-spintronics,skyrmions,spintronics,unconventional computing},
+ file = {/Users/wasmer/Nextcloud/Zotero/Finocchio et al_2024_Roadmap for unconventional computing with nanotechnology.pdf}
+}
+
@online{finziPracticalMethodConstructing2021,
title = {A {{Practical Method}} for {{Constructing Equivariant Multilayer Perceptrons}} for {{Arbitrary Matrix Groups}}},
author = {Finzi, Marc and Welling, Max and Wilson, Andrew Gordon},
date = {2021-04-19},
eprint = {2104.09459},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math, stat},
doi = {10.48550/arXiv.2104.09459},
url = {http://arxiv.org/abs/2104.09459},
urldate = {2023-08-22},
abstract = {Symmetries and equivariance are fundamental to the generalization of neural networks on domains such as images, graphs, and point clouds. Existing work has primarily focused on a small number of groups, such as the translation, rotation, and permutation groups. In this work we provide a completely general algorithm for solving for the equivariant layers of matrix groups. In addition to recovering solutions from other works as special cases, we construct multilayer perceptrons equivariant to multiple groups that have never been tackled before, including \$\textbackslash mathrm\{O\}(1,3)\$, \$\textbackslash mathrm\{O\}(5)\$, \$\textbackslash mathrm\{Sp\}(n)\$, and the Rubik's cube group. Our approach outperforms non-equivariant baselines, with applications to particle physics and dynamical systems. We release our software library to enable researchers to construct equivariant layers for arbitrary matrix groups.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Finzi et al_2021_A Practical Method for Constructing Equivariant Multilayer Perceptrons for.pdf;/Users/wasmer/Zotero/storage/CFQ89M8L/2104.html}
}
+@online{fisherMultitaskMethodsPredicting2024,
+ title = {Multitask Methods for Predicting Molecular Properties from Heterogeneous Data},
+ author = {Fisher, Katharine and Herbst, Michael and Marzouk, Youssef},
+ date = {2024-05-24},
+ eprint = {2401.17898},
+ eprinttype = {arXiv},
+ eprintclass = {physics, stat},
+ doi = {10.48550/arXiv.2401.17898},
+ url = {http://arxiv.org/abs/2401.17898},
+ urldate = {2024-06-21},
+ abstract = {Data generation remains a bottleneck in training surrogate models to predict molecular properties. We demonstrate that multitask Gaussian process regression overcomes this limitation by leveraging both expensive and cheap data sources. In particular, we consider training sets constructed from coupled-cluster (CC) and density functional theory (DFT) data. We report that multitask surrogates can predict at CC-level accuracy with a reduction to data generation cost by over an order of magnitude. Of note, our approach allows the training set to include DFT data generated by a heterogeneous mix of exchange-correlation functionals without imposing any artificial hierarchy on functional accuracy. More generally, the multitask framework can accommodate a wider range of training set structures -- including full disparity between the different levels of fidelity -- than existing kernel approaches based on \$\textbackslash Delta\$-learning, though we show that the accuracy of the two approaches can be similar. Consequently, multitask regression can be a tool for reducing data generation costs even further by opportunistically exploiting existing data sources.},
+ pubstate = {prepublished},
+ keywords = {AML,coupled cluster,DFT,Gaussian process,GPR,ML,ML-ESM,multi-fidelity,multi-task learning,multimodal input,multitask learning},
+ file = {/Users/wasmer/Nextcloud/Zotero/Fisher et al_2024_Multitask methods for predicting molecular properties from heterogeneous data.pdf;/Users/wasmer/Zotero/storage/P5Z73WWE/2401.html}
+}
+
@book{fleuretLittleBookDeep2023,
title = {The {{Little Book}} of {{Deep Learning}}},
author = {Fleuret, François},
@@ -5150,7 +5643,7 @@ Junqi Yin\\
abstract = {This book is a short introduction to deep learning for readers with a STEM background, originally designed to be read on a phone screen. It is distributed under a non-commercial Creative Commons license and was downloaded close to 250'000 times in the month following its public release.},
langid = {english},
pagetotal = {168},
- keywords = {\_tablet,Deep learning,educational,General ML,learning material,ML theory,online book,textbook},
+ keywords = {Deep learning,educational,General ML,learning material,ML theory,online book,textbook},
file = {/Users/wasmer/Nextcloud/Zotero/Fleuret_2023_The Little Book of Deep Learning.pdf}
}
@@ -5176,13 +5669,13 @@ Junqi Yin\\
author = {Focassio, Bruno and Domina, Michelangelo and Patil, Urvesh and Fazzio, Adalberto and Sanvito, Stefano},
date = {2023-01-31},
eprint = {2301.13550},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2301.13550},
url = {http://arxiv.org/abs/2301.13550},
urldate = {2023-02-23},
abstract = {As the go-to method to solve the electronic structure problem, Kohn-Sham density functional theory (KS-DFT) can be used to obtain the ground-state charge density, total energy, and several other key materials' properties. Unfortunately, the solution of the Kohn-Sham equations is found iteratively. This is a numerically intensive task, limiting the possible size and complexity of the systems to be treated. Machine-learning (ML) models for the charge density can then be used as surrogates to generate the converged charge density and reduce the computational cost of solving the electronic structure problem. We derive a powerful grid-centred structural representation based on the Jacobi and Legendre polynomials that, combined with a linear regression built on a data-efficient workflow, can accurately learn the charge density. Then, we design a machine-learning pipeline that can return energy and forces at the quality of a converged DFT calculation but at a fraction of the computational cost. This can be used as a tool for the fast scanning of the energy landscape and as a starting point to the DFT self-consistent cycle, in both cases maintaining a low computational cost.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {\_tablet,AML,DFT,grid-based descriptors,Jacobi-Legendre,library,ML,ML-Density,ML-DFT,ML-ESM,prediction of electron density,VASP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Focassio et al_2023_Linear Jacobi-Legendre expansion of the charge density for machine.pdf;/Users/wasmer/Zotero/storage/HPSZ89R2/2301.html}
}
@@ -5204,8 +5697,8 @@ Junqi Yin\\
abstract = {Kohn–Sham density functional theory (KS-DFT) is a powerful method to obtain key materials’ properties, but the iterative solution of the KS equations is a numerically intensive task, which limits its application to complex systems. To address this issue, machine learning (ML) models can be used as surrogates to find the ground-state charge density and reduce the computational overheads. We develop a grid-centred structural representation, based on Jacobi and Legendre polynomials combined with a linear regression, to accurately learn the converged DFT charge density. This integrates into a ML pipeline that can return any density-dependent observable, including energy and forces, at the quality of a converged DFT calculation, but at a fraction of the computational cost. Fast scanning of energy landscapes and producing starting densities for the DFT self-consistent cycle are among the applications of our scheme.},
issue = {1},
langid = {english},
- keywords = {\_tablet,AML,DFT,grid-based descriptors,Jacobi-Legendre,library,ML,ML-Density,ML-DFT,ML-ESM,prediction of electron density,VASP,with-code},
- file = {/Users/wasmer/Nextcloud/Zotero/Focassio et al_2023_Linear Jacobi-Legendre expansion of the charge density for machine2.pdf}
+ keywords = {AML,DFT,grid-based descriptors,Jacobi-Legendre,library,ML,ML-Density,ML-DFT,ML-ESM,prediction of electron density,VASP,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Focassio et al_2023_Linear Jacobi-Legendre expansion of the charge density for machine2.pdf;/Users/wasmer/Nextcloud/Zotero/Focassio et al_2023_Linear Jacobi-Legendre expansion of the charge density for machine3.pdf}
}
@online{focassioPerformanceAssessmentUniversal2024,
@@ -5214,14 +5707,14 @@ Junqi Yin\\
author = {Focassio, Bruno and Freitas, Luis Paulo Mezzina and Schleder, Gabriel R.},
date = {2024-03-06},
eprint = {2403.04217},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2403.04217},
url = {http://arxiv.org/abs/2403.04217},
urldate = {2024-03-14},
abstract = {Machine learning interatomic potentials (MLIPs) are one of the main techniques in the materials science toolbox, able to bridge ab initio accuracy with the computational efficiency of classical force fields. This allows simulations ranging from atoms, molecules, and biosystems, to solid and bulk materials, surfaces, nanomaterials, and their interfaces and complex interactions. A recent class of advanced MLIPs, which use equivariant representations and deep graph neural networks, is known as universal models. These models are proposed as foundational models suitable for any system, covering most elements from the periodic table. Current universal MLIPs (UIPs) have been trained with the largest consistent dataset available nowadays. However, these are composed mostly of bulk materials' DFT calculations. In this article, we assess the universality of all openly available UIPs, namely MACE, CHGNet, and M3GNet, in a representative task of generalization: calculation of surface energies. We find that the out-of-the-box foundational models have significant shortcomings in this task, with errors correlated to the total energy of surface simulations, having an out-of-domain distance from the training dataset. Our results show that while UIPs are an efficient starting point for fine-tuning specialized models, we envision the potential of increasing the coverage of the materials space towards universal training datasets for MLIPs.},
- pubstate = {preprint},
- keywords = {\_tablet,AML,benchmarking,CHGNet,disordered,foundation models,M3GNet,MACE,materials project,ML,MLP,MLP comparison,MTP,NequIP,surface physics,todo-tagging,universal potential},
+ pubstate = {prepublished},
+ keywords = {AML,benchmarking,CHGNet,disordered,foundation models,M3GNet,MACE,materials project,ML,MLP,MLP comparison,MTP,NequIP,surface physics,todo-tagging,universal potential},
file = {/Users/wasmer/Nextcloud/Zotero/Focassio et al_2024_Performance Assessment of Universal Machine Learning Interatomic Potentials.pdf;/Users/wasmer/Zotero/storage/QLASD4BQ/2403.html}
}
@@ -5239,7 +5732,7 @@ Junqi Yin\\
eventtitle = {Autumn {{School}} on {{Correlated Electrons}}},
isbn = {9783958064669},
langid = {english},
- keywords = {\_tablet,ACE,cusps,FermiNet,ML-ESM,ML-QMBP,prediction of wavefunction,QMC,Slater-Jastrow,VMC},
+ keywords = {ACE,cusps,FermiNet,ML-ESM,ML-QMBP,prediction of wavefunction,QMC,Slater-Jastrow,VMC},
file = {/Users/wasmer/Nextcloud/Zotero/Foulkes_Drautz_2020_Topology, Entanglement, and Strong Correlations.pdf;/Users/wasmer/Zotero/storage/WLIE37SZ/884084.html}
}
@@ -5253,7 +5746,7 @@ Junqi Yin\\
doi = {10.1109/IJCNN55064.2022.9892358},
abstract = {Learning 3D representations of point clouds that generalize well to arbitrary orientations is a challenge of practical importance in domains ranging from computer vision to molecular modeling. The proposed approach uses a concentric spherical spatial representation, formed by nesting spheres discretized the icosahedral grid, as the basis for structured learning over point clouds. We propose rotationally equivariant convolutions for learning over the concentric spherical grid, which are incorporated into a novel architecture for representation learning that is robust to general rotations in 3D. We demonstrate the effectiveness and extensibility of our approach to problems in different domains, such as 3D shape recognition and predicting fundamental properties of molecular systems.},
eventtitle = {2022 {{International Joint Conference}} on {{Neural Networks}} ({{IJCNN}})},
- keywords = {\_tablet,AML,convolution,CSNN,equivariant,General ML,GNN,library,ML,ML-DFT,ML-ESM,NN,original publication,point cloud compression,point cloud data,prediction of DOS,representation learning,spherical convolution,spherical NN},
+ keywords = {AML,convolution,CSNN,equivariant,General ML,GNN,library,ML,ML-DFT,ML-ESM,NN,original publication,point cloud compression,point cloud data,prediction of DOS,representation learning,spherical convolution,spherical NN},
file = {/Users/wasmer/Nextcloud/Zotero/Fox et al_2022_Concentric Spherical Neural Network for 3D Representation Learning.pdf;/Users/wasmer/Zotero/storage/YBYXAFZ3/9892358.html}
}
@@ -5263,7 +5756,7 @@ Junqi Yin\\
author = {Frank, Thorben and Chmiela, Stefan},
date = {2021-09-06},
eprint = {2106.02549},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
publisher = {arXiv},
doi = {10.48550/arXiv.2106.02549},
@@ -5280,14 +5773,14 @@ Junqi Yin\\
author = {Frank, J. Thorben and Unke, Oliver T. and Müller, Klaus-Robert},
date = {2023-01-09},
eprint = {2205.14276},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2205.14276},
url = {http://arxiv.org/abs/2205.14276},
urldate = {2023-04-04},
abstract = {The application of machine learning methods in quantum chemistry has enabled the study of numerous chemical phenomena, which are computationally intractable with traditional ab-initio methods. However, some quantum mechanical properties of molecules and materials depend on non-local electronic effects, which are often neglected due to the difficulty of modeling them efficiently. This work proposes a modified attention mechanism adapted to the underlying physics, which allows to recover the relevant non-local effects. Namely, we introduce spherical harmonic coordinates (SPHCs) to reflect higher-order geometric information for each atom in a molecule, enabling a non-local formulation of attention in the SPHC space. Our proposed model So3krates - a self-attention based message passing neural network - uncouples geometric information from atomic features, making them independently amenable to attention mechanisms. Thereby we construct spherical filters, which extend the concept of continuous filters in Euclidean space to SPHC space and serve as foundation for a spherical self-attention mechanism. We show that in contrast to other published methods, So3krates is able to describe non-local quantum mechanical effects over arbitrary length scales. Further, we find evidence that the inclusion of higher-order geometric correlations increases data efficiency and improves generalization. So3krates matches or exceeds state-of-the-art performance on popular benchmarks, notably, requiring a significantly lower number of parameters (0.25 - 0.4x) while at the same time giving a substantial speedup (6 - 14x for training and 2 - 11x for inference) compared to other models.},
- pubstate = {preprint},
- keywords = {\_tablet,AML,attention,equivariant,Euclidean space,FLAX,GDL,invariance,JAX,library,long-range interaction,ML,ML-FF,molecules,MPNN,NequIP,original publication,QM7-X,SchNet,sGDML,SO(3),spherical harmonic coordinates,SpookyNet,with-code},
+ pubstate = {prepublished},
+ keywords = {AML,attention,equivariant,Euclidean space,FLAX,GDL,invariance,JAX,library,long-range interaction,ML,ML-FF,molecules,MPNN,NequIP,original publication,QM7-X,SchNet,sGDML,SO(3),spherical harmonic coordinates,SpookyNet,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Frank et al_2023_So3krates.pdf;/Users/wasmer/Zotero/storage/9GZJ7VNS/2205.html}
}
@@ -5306,7 +5799,7 @@ Junqi Yin\\
urldate = {2022-10-03},
abstract = {Fraux et al., (2020). Chemiscope: interactive structure-property explorer for materials and molecules. Journal of Open Source Software, 5(51), 2117, https://doi.org/10.21105/joss.02117},
langid = {english},
- keywords = {\_tablet,data exploration,Database,ML,molecules,sketchmap,solids,unsupervised learning,visualization},
+ keywords = {data exploration,Database,ML,molecules,sketchmap,solids,unsupervised learning,visualization},
file = {/Users/wasmer/Nextcloud/Zotero/Fraux et al_2020_Chemiscope.pdf;/Users/wasmer/Zotero/storage/TCQI9XE2/joss.html}
}
@@ -5325,7 +5818,7 @@ Junqi Yin\\
url = {https://www.sciencedirect.com/science/article/pii/S2666389921001884},
urldate = {2023-09-19},
abstract = {In this paper, we critique ICT's current and projected climate impacts. Peer-reviewed studies estimate ICT's current share of global greenhouse gas (GHG) emissions at 1.8\%–2.8\% of global GHG emissions; adjusting for truncation of supply chain pathways, we find that this share could actually be between 2.1\% and 3.9\%. For ICT's future emissions, we explore assumptions underlying analysts' projections to understand the reasons for their variability. All analysts agree that ICT emissions will not reduce without major concerted efforts involving broad political and industrial action. We provide three reasons to believe ICT emissions are going to increase barring intervention and find that not all carbon pledges in the ICT sector are ambitious enough to meet climate targets. We explore the underdevelopment of policy mechanisms for enforcing sector-wide compliance, and contend that, without a global carbon constraint, a new regulatory framework is required to keep the ICT sector's footprint aligned with the Paris Agreement.},
- keywords = {/unread,ecological footprint,economics,energy consumption,energy efficiency,environmental impact,for introductions,ICT sector,low-power electronics,world energy consumption},
+ keywords = {ecological footprint,economics,energy consumption,energy efficiency,environmental impact,for introductions,ICT sector,low-power electronics,world energy consumption},
file = {/Users/wasmer/Nextcloud/Zotero/Freitag et al_2021_The real climate and transformative impact of ICT.pdf;/Users/wasmer/Zotero/storage/3IPQYR9I/S2666389921001884.html}
}
@@ -5389,7 +5882,7 @@ Junqi Yin\\
url = {https://link.aps.org/doi/10.1103/PhysRevB.105.014103},
urldate = {2023-05-06},
abstract = {The commonly employed supercell approach for defects in crystalline materials may introduce spurious interactions between the defect and its periodic images. A rich literature is available on how the interaction energies can be estimated, reduced, or corrected. A simple and seemingly straightforward approach is to extrapolate from a series of finite supercell sizes to the infinite-size limit, assuming a smooth polynomial dependence of the energy on inverse supercell size. In this work, we demonstrate by means of explict density-functional theory supercell calculations and simplified models that wave-function overlap and electrostatic interactions lead to more complex dependencies on supercell size than commonly assumed. We show that this complexity cannot be captured by the simple extrapolation approaches and that suitable correction schemes should be employed.},
- keywords = {/unread,\_tablet,2D material,defects,DFT,impurity embedding,interfaces and thin films,KKR,point defects,supercell,surface physics},
+ keywords = {/unread,2D material,defects,DFT,impurity embedding,interfaces and thin films,KKR,point defects,supercell,surface physics},
file = {/Users/wasmer/Nextcloud/Zotero/Freysoldt et al_2022_Limitations of empirical supercell extrapolation for calculations of point.pdf;/Users/wasmer/Zotero/storage/EAVNV5DB/PhysRevB.105.html}
}
@@ -5442,13 +5935,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Frolov, Sergey and Mourik, Vincent},
date = {2022-03-31},
eprint = {2203.17060},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:quant-ph},
doi = {10.48550/arXiv.2203.17060},
url = {http://arxiv.org/abs/2203.17060},
urldate = {2023-09-18},
abstract = {In 2011-2012 we performed experiments on hybrid superconductor-semiconductor nanowire devices which yielded signatures of Majorana fermions based on zero-bias peaks in tunneling measurements. The research field that grew out of those findings and other contemporary works has advanced significantly, and a lot of new knowledge and insights were gained. However, key smoking gun evidence of Majorana is still lacking. In this paper, we report that while reviewing our old data recently, armed with a decade of knowledge, we realized that back in 2012 our results contained two breakthrough Majorana discoveries. Specifically, we have observed quantized zero-bias peaks, the hallmark of ideal Majorana states. Furthermore, we have observed the closing and re-opening of the induced gap perfectly correlated with the emergence of the zero-bias peak - clear evidence of the topological quantum phase superconducting transition. These insights should pave the way to topological Majorana qubits, and you should also check supplementary information for important disclosures.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,experimental,failure,FZJ,Majorana,MZM,PGI,physics,superconductor,topological,Topological Superconductor},
file = {/Users/wasmer/Nextcloud/Zotero/Frolov_Mourik_2022_We cannot believe we overlooked these Majorana discoveries.pdf;/Users/wasmer/Zotero/storage/IR3K8NZ9/2203.html}
}
@@ -5459,13 +5952,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Fuchs, Fabian B. and Worrall, Daniel E. and Fischer, Volker and Welling, Max},
date = {2020-11-24},
eprint = {2006.10503},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.2006.10503},
url = {http://arxiv.org/abs/2006.10503},
urldate = {2022-10-03},
abstract = {We introduce the SE(3)-Transformer, a variant of the self-attention module for 3D point clouds and graphs, which is equivariant under continuous 3D roto-translations. Equivariance is important to ensure stable and predictable performance in the presence of nuisance transformations of the data input. A positive corollary of equivariance is increased weight-tying within the model. The SE(3)-Transformer leverages the benefits of self-attention to operate on large point clouds and graphs with varying number of points, while guaranteeing SE(3)-equivariance for robustness. We evaluate our model on a toy N-body particle simulation dataset, showcasing the robustness of the predictions under rotations of the input. We further achieve competitive performance on two real-world datasets, ScanObjectNN and QM9. In all cases, our model outperforms a strong, non-equivariant attention baseline and an equivariant model without attention.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {equivariant,GCN,general ML,GNN,library,ML,QM9,SchNet,SE(3),self-attention,transformer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Fuchs et al_2020_SE(3)-Transformers.pdf;/Users/wasmer/Zotero/storage/UMVV286P/2006.html}
}
@@ -5476,13 +5969,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Fu, Xiang and Wu, Zhenghao and Wang, Wujie and Xie, Tian and Keten, Sinan and Gomez-Bombarelli, Rafael and Jaakkola, Tommi},
date = {2022-10-13},
eprint = {2210.07237},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2210.07237},
url = {http://arxiv.org/abs/2210.07237},
urldate = {2023-04-03},
abstract = {Molecular dynamics (MD) simulation techniques are widely used for various natural science applications. Increasingly, machine learning (ML) force field (FF) models begin to replace ab-initio simulations by predicting forces directly from atomic structures. Despite significant progress in this area, such techniques are primarily benchmarked by their force/energy prediction errors, even though the practical use case would be to produce realistic MD trajectories. We aim to fill this gap by introducing a novel benchmark suite for ML MD simulation. We curate representative MD systems, including water, organic molecules, peptide, and materials, and design evaluation metrics corresponding to the scientific objectives of respective systems. We benchmark a collection of state-of-the-art (SOTA) ML FF models and illustrate, in particular, how the commonly benchmarked force accuracy is not well aligned with relevant simulation metrics. We demonstrate when and how selected SOTA methods fail, along with offering directions for further improvement. Specifically, we identify stability as a key metric for ML models to improve. Our benchmark suite comes with a comprehensive open-source codebase for training and simulation with ML FFs to facilitate further work.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,benchmarking,DimeNet,equivariant,GemNet,invariance,MD,MD17,Microsoft Research,ML,ML-FF,MLP,NequIP,PES,SchNet},
file = {/Users/wasmer/Nextcloud/Zotero/Fu et al_2022_Forces are not Enough.pdf;/Users/wasmer/Zotero/storage/2GZERUJD/2210.html}
}
@@ -5501,7 +5994,7 @@ Subject\_term: Quantum physics, Publishing, Peer review},
url = {https://link.aps.org/doi/10.1103/PhysRevMaterials.6.023802},
urldate = {2023-04-04},
abstract = {Automatic exhaustive exploration of a large material space by high-performance supercomputers is crucial for developing new functional materials. We demonstrated the efficiency of high-throughput calculations using the all-electron Korringa-Kohn-Rostoker coherent potential approximation method with the density functional theory for the large material space consisting of quaternary high entropy alloys, which are nonstoichiometric and substitutionally disordered materials. The exhaustive calculations were performed for 147 630 systems based on the AkaiKKR program package and supercomputer Fugaku, where the numerical parameters and self-consistent convergence are automatically controlled. The large material database including the total energies, magnetization, Curie temperature, and residual resistivity was constructed by our calculations. We used frequent itemset mining to identify the characteristics of parcels in magnetization and Curie temperature space. We also identified the elements that enhance the magnetization and Curie temperature and clarified the rough dependence of the elements through regression modeling of the residual resistivity.},
- keywords = {/unread,\_tablet,CPA,HTC,KKR},
+ keywords = {/unread,CPA,HTC,KKR},
file = {/Users/wasmer/Nextcloud/Zotero/Fukushima et al_2022_Automatic exhaustive calculations of large material space by.pdf;/Users/wasmer/Zotero/storage/VNUQ6LGT/PhysRevMaterials.6.html}
}
@@ -5596,7 +6089,7 @@ Subject\_term: Quantum physics, Publishing, Peer review},
urldate = {2023-05-20},
abstract = {We present a model-agnostic method that gives natural language explanations of molecular structure property predictions. Machine learning models are now common for molecular property prediction and chemical design. They typically are black boxes -- having no explanation for predictions. We show how to use surrogate models to attribute predictions to chemical descriptors and molecular substructures, independent of the black box model inputs. The method generates explanations consistent with chemical reasoning, like connecting existence of a functional group or molecular polarity. We see in a genuine test like blood brain barrier permeation, our descriptor explanations match biologically observed SARs with mechanistic support. We show these quantitative explanations can be further translated to natural language.},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,chemistry,contrastive explanation,counterfactual explanation,Deep learning,descriptor explanation,descriptors,descriptors analysis,feature attribution,GPT,GPT-3,graph attribution,library,LIME,LLM,MACCS,ML,model explanation,nlp,property prediction,QSAR,SAR,SELFIES,SHAP,STONED,surrogate model,with-code,XAI},
file = {/Users/wasmer/Nextcloud/Zotero/Gandhi_White_2022_Explaining molecular properties with natural language.pdf}
}
@@ -5622,22 +6115,42 @@ Subject\_term: Quantum physics, Publishing, Peer review},
file = {/Users/wasmer/Nextcloud/Zotero/Gao_Remsing_2022_Self-consistent determination of long-range electrostatics in neural network.pdf}
}
-@online{gardnerSyntheticDataEnable2022a,
+@online{gardnerSyntheticDataEnable2022,
title = {Synthetic Data Enable Experiments in Atomistic Machine Learning},
author = {Gardner, John L. A. and Beaulieu, Zoé Faure and Deringer, Volker L.},
date = {2022-11-29},
eprint = {2211.16443},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2211.16443},
url = {http://arxiv.org/abs/2211.16443},
urldate = {2023-03-01},
abstract = {Machine-learning models are increasingly used to predict properties of atoms in chemical systems. There have been major advances in developing descriptors and regression frameworks for this task, typically starting from (relatively) small sets of quantum-mechanical reference data. Larger datasets of this kind are becoming available, but remain expensive to generate. Here we demonstrate the use of a large dataset that we have "synthetically" labelled with per-atom energies from an existing ML potential model. The cheapness of this process, compared to the quantum-mechanical ground truth, allows us to generate millions of datapoints, in turn enabling rapid experimentation with atomistic ML models from the small- to the large-data regime. This approach allows us here to compare regression frameworks in depth, and to explore visualisation based on learned representations. We also show that learning synthetic data labels can be a useful pre-training task for subsequent fine-tuning on small datasets. In the future, we expect that our open-sourced dataset, and similar ones, will be useful in rapidly exploring deep-learning models in the limit of abundant chemical data.},
- pubstate = {preprint},
- keywords = {/unread,carbon,data augmentation,Database,disordered,DKL,GAP,GPR,KPCovR,LAMMPS,MD,ML,MLP,NN,prediction of potential energy,small data,SOAP,Supervised learning,synthetic data,UMAP,unsupervised learning},
+ pubstate = {prepublished},
+ keywords = {carbon,data augmentation,Database,disordered,DKL,GAP,GPR,KPCovR,LAMMPS,MD,ML,MLP,NN,prediction of potential energy,small data,SOAP,Supervised learning,synthetic data,UMAP,unsupervised learning},
file = {/Users/wasmer/Nextcloud/Zotero/Gardner et al_2022_Synthetic data enable experiments in atomistic machine learning2.pdf;/Users/wasmer/Zotero/storage/99FPUBGW/2211.html}
}
+@article{gardnerSyntheticPretrainingNeuralnetwork2024,
+ title = {Synthetic Pre-Training for Neural-Network Interatomic Potentials},
+ author = {Gardner, John L. A. and Baker, Kathryn T. and Deringer, Volker L.},
+ date = {2024-01},
+ journaltitle = {Machine Learning: Science and Technology},
+ shortjournal = {Mach. Learn.: Sci. Technol.},
+ volume = {5},
+ number = {1},
+ pages = {015003},
+ publisher = {IOP Publishing},
+ issn = {2632-2153},
+ doi = {10.1088/2632-2153/ad1626},
+ url = {https://dx.doi.org/10.1088/2632-2153/ad1626},
+ urldate = {2024-06-14},
+ abstract = {Machine learning (ML) based interatomic potentials have transformed the field of atomistic materials modelling. However, ML potentials depend critically on the quality and quantity of quantum-mechanical reference data with which they are trained, and therefore developing datasets and training pipelines is becoming an increasingly central challenge. Leveraging the idea of ‘synthetic’ (artificial) data that is common in other areas of ML research, we here show that synthetic atomistic data, themselves obtained at scale with an existing ML potential, constitute a useful pre-training task for neural-network (NN) interatomic potential models. Once pre-trained with a large synthetic dataset, these models can be fine-tuned on a much smaller, quantum-mechanical one, improving numerical accuracy and stability in computational practice. We demonstrate feasibility for a series of equivariant graph-NN potentials for carbon, and we carry out initial experiments to test the limits of the approach.},
+ langid = {english},
+ keywords = {/unread,ACE,AML,data augmentation,database generation,fine-tuning,GAP,GNN,library,MD,ML,MLP,NequIP,pretrained models,pretraining,synthetic data,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Gardner et al_2024_Synthetic pre-training for neural-network interatomic potentials.pdf}
+}
+
@article{garridotorresLowScalingAlgorithmNudged2019,
title = {Low-{{Scaling Algorithm}} for {{Nudged Elastic Band Calculations Using}} a {{Surrogate Machine Learning Model}}},
author = {Garrido Torres, José A. and Jennings, Paul C. and Hansen, Martin H. and Boes, Jacob R. and Bligaard, Thomas},
@@ -5677,13 +6190,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Gasteiger, Johannes and Groß, Janek and Günnemann, Stephan},
date = {2022-04-05},
eprint = {2003.03123},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2003.03123},
url = {http://arxiv.org/abs/2003.03123},
urldate = {2022-10-03},
abstract = {Graph neural networks have recently achieved great successes in predicting quantum mechanical properties of molecules. These models represent a molecule as a graph using only the distance between atoms (nodes). They do not, however, consider the spatial direction from one atom to another, despite directional information playing a central role in empirical potentials for molecules, e.g. in angular potentials. To alleviate this limitation we propose directional message passing, in which we embed the messages passed between atoms instead of the atoms themselves. Each message is associated with a direction in coordinate space. These directional message embeddings are rotationally equivariant since the associated directions rotate with the molecule. We propose a message passing scheme analogous to belief propagation, which uses the directional information by transforming messages based on the angle between them. Additionally, we use spherical Bessel functions and spherical harmonics to construct theoretically well-founded, orthogonal representations that achieve better performance than the currently prevalent Gaussian radial basis representations while using fewer than 1/4 of the parameters. We leverage these innovations to construct the directional message passing neural network (DimeNet). DimeNet outperforms previous GNNs on average by 76\% on MD17 and by 31\% on QM9. Our implementation is available online.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DimeNet,GNN,MD,ML,MLP,molecules,MPNN,original publication,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Gasteiger et al_2022_Directional Message Passing for Molecular Graphs.pdf;/Users/wasmer/Zotero/storage/G7KWBFCS/2003.html}
}
@@ -5693,13 +6206,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Gasteiger, Johannes and Yeshwanth, Chandan and Günnemann, Stephan},
date = {2022-04-05},
eprint = {2111.04718},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, q-bio},
doi = {10.48550/arXiv.2111.04718},
url = {http://arxiv.org/abs/2111.04718},
urldate = {2022-10-03},
abstract = {Graph neural networks that leverage coordinates via directional message passing have recently set the state of the art on multiple molecular property prediction tasks. However, they rely on atom position information that is often unavailable, and obtaining it is usually prohibitively expensive or even impossible. In this paper we propose synthetic coordinates that enable the use of advanced GNNs without requiring the true molecular configuration. We propose two distances as synthetic coordinates: Distance bounds that specify the rough range of molecular configurations, and graph-based distances using a symmetric variant of personalized PageRank. To leverage both distance and angular information we propose a method of transforming normal graph neural networks into directional MPNNs. We show that with this transformation we can reduce the error of a normal graph neural network by 55\% on the ZINC benchmark. We furthermore set the state of the art on ZINC and coordinate-free QM9 by incorporating synthetic coordinates in the SMP and DimeNet++ models. Our implementation is available online.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DimeNet,DimeNet++,GNN,MD,MLP,molecules,MPNN,open-review},
file = {/Users/wasmer/Nextcloud/Zotero/Gasteiger et al_2022_Directional Message Passing on Molecular Graphs via Synthetic Coordinates.pdf;/Users/wasmer/Zotero/storage/FEPN4JW4/2111.html}
}
@@ -5709,13 +6222,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Gasteiger, Johannes and Giri, Shankari and Margraf, Johannes T. and Günnemann, Stephan},
date = {2022-04-05},
eprint = {2011.14115},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2011.14115},
url = {http://arxiv.org/abs/2011.14115},
urldate = {2022-10-03},
abstract = {Many important tasks in chemistry revolve around molecules during reactions. This requires predictions far from the equilibrium, while most recent work in machine learning for molecules has been focused on equilibrium or near-equilibrium states. In this paper we aim to extend this scope in three ways. First, we propose the DimeNet++ model, which is 8x faster and 10\% more accurate than the original DimeNet on the QM9 benchmark of equilibrium molecules. Second, we validate DimeNet++ on highly reactive molecules by developing the challenging COLL dataset, which contains distorted configurations of small molecules during collisions. Finally, we investigate ensembling and mean-variance estimation for uncertainty quantification with the goal of accelerating the exploration of the vast space of non-equilibrium structures. Our DimeNet++ implementation as well as the COLL dataset are available online.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DimeNet,DimeNet++,GNN,MD,ML,MLP,molecules,MPNN,original publication,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Gasteiger et al_2022_Fast and Uncertainty-Aware Directional Message Passing for Non-Equilibrium.pdf;/Users/wasmer/Zotero/storage/BVEXST79/2011.html}
}
@@ -5726,13 +6239,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Gasteiger, Johannes and Becker, Florian and Günnemann, Stephan},
date = {2022-04-05},
eprint = {2106.08903},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2106.08903},
url = {http://arxiv.org/abs/2106.08903},
urldate = {2022-10-03},
abstract = {Effectively predicting molecular interactions has the potential to accelerate molecular dynamics by multiple orders of magnitude and thus revolutionize chemical simulations. Graph neural networks (GNNs) have recently shown great successes for this task, overtaking classical methods based on fixed molecular kernels. However, they still appear very limited from a theoretical perspective, since regular GNNs cannot distinguish certain types of graphs. In this work we close this gap between theory and practice. We show that GNNs with directed edge embeddings and two-hop message passing are indeed universal approximators for predictions that are invariant to translation, and equivariant to permutation and rotation. We then leverage these insights and multiple structural improvements to propose the geometric message passing neural network (GemNet). We demonstrate the benefits of the proposed changes in multiple ablation studies. GemNet outperforms previous models on the COLL, MD17, and OC20 datasets by 34\%, 41\%, and 20\%, respectively, and performs especially well on the most challenging molecules. Our implementation is available online.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DimeNet,DimeNet++,GemNet,MD,ML,MLP,molecules,MPNN,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Gasteiger et al_2022_GemNet.pdf;/Users/wasmer/Zotero/storage/FE5R77B9/2106.html}
}
@@ -5742,13 +6255,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Gavini, Vikram and Baroni, Stefano and Blum, Volker and Bowler, David R. and Buccheri, Alexander and Chelikowsky, James R. and Das, Sambit and Dawson, William and Delugas, Pietro and Dogan, Mehmet and Draxl, Claudia and Galli, Giulia and Genovese, Luigi and Giannozzi, Paolo and Giantomassi, Matteo and Gonze, Xavier and Govoni, Marco and Gulans, Andris and Gygi, François and Herbert, John M. and Kokott, Sebastian and Kühne, Thomas D. and Liou, Kai-Hsin and Miyazaki, Tsuyoshi and Motamarri, Phani and Nakata, Ayako and Pask, John E. and Plessl, Christian and Ratcliff, Laura E. and Richard, Ryan M. and Rossi, Mariana and Schade, Robert and Scheffler, Matthias and Schütt, Ole and Suryanarayana, Phanish and Torrent, Marc and Truflandier, Lionel and Windus, Theresa L. and Xu, Qimen and Yu, Victor W.-Z. and Perez, Danny},
date = {2022-09-26},
eprint = {2209.12747},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2209.12747},
url = {http://arxiv.org/abs/2209.12747},
urldate = {2023-06-30},
abstract = {Electronic structure calculations have been instrumental in providing many important insights into a range of physical and chemical properties of various molecular and solid-state systems. Their importance to various fields, including materials science, chemical sciences, computational chemistry and device physics, is underscored by the large fraction of available public supercomputing resources devoted to these calculations. As we enter the exascale era, exciting new opportunities to increase simulation numbers, sizes, and accuracies present themselves. In order to realize these promises, the community of electronic structure software developers will however first have to tackle a number of challenges pertaining to the efficient use of new architectures that will rely heavily on massive parallelism and hardware accelerators. This roadmap provides a broad overview of the state-of-the-art in electronic structure calculations and of the various new directions being pursued by the community. It covers 14 electronic structure codes, presenting their current status, their development priorities over the next five years, and their plans towards tackling the challenges and leveraging the opportunities presented by the advent of exascale computing.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ABINIT,BigDFT,chemistry,CONQUEST,CP2K,DFT,DFT-FE,ESM,exascale,FHI-aims,HPC,ML-DFT,NWChem,PARSEC,physics,Q-Chem,Quantum ESPRESSO,roadmap},
file = {/Users/wasmer/Nextcloud/Zotero/Gavini et al_2022_Roadmap on Electronic Structure Codes in the Exascale Era.pdf;/Users/wasmer/Zotero/storage/B2ZN8AD8/2209.html}
}
@@ -5799,14 +6312,14 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Geiger, Mario and Smidt, Tess},
date = {2022-07-18},
eprint = {2207.09453},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2207.09453},
url = {http://arxiv.org/abs/2207.09453},
urldate = {2022-08-21},
abstract = {We present e3nn, a generalized framework for creating E(3) equivariant trainable functions, also known as Euclidean neural networks. e3nn naturally operates on geometry and geometric tensors that describe systems in 3D and transform predictably under a change of coordinate system. The core of e3nn are equivariant operations such as the TensorProduct class or the spherical harmonics functions that can be composed to create more complex modules such as convolutions and attention mechanisms. These core operations of e3nn can be used to efficiently articulate Tensor Field Networks, 3D Steerable CNNs, Clebsch-Gordan Networks, SE(3) Transformers and other E(3) equivariant networks.},
- pubstate = {preprint},
- keywords = {\_tablet,e3nn,EGNN,ENN,equivariant,library,ML-ESM,prediction of electron density,with-code},
+ pubstate = {prepublished},
+ keywords = {e3nn,EGNN,ENN,equivariant,library,ML-ESM,prediction of electron density,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Geiger_Smidt_2022_e3nn.pdf;/Users/wasmer/Zotero/storage/SJW8392C/2207.html}
}
@@ -5843,13 +6356,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Gerard, Leon and Scherbela, Michael and Marquetand, Philipp and Grohs, Philipp},
date = {2022-05-31},
eprint = {2205.09438},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2205.09438},
url = {http://arxiv.org/abs/2205.09438},
urldate = {2022-08-16},
abstract = {Finding accurate solutions to the Schr\textbackslash "odinger equation is the key unsolved challenge of computational chemistry. Given its importance for the development of new chemical compounds, decades of research have been dedicated to this problem, but due to the large dimensionality even the best available methods do not yet reach the desired accuracy. Recently the combination of deep learning with Monte Carlo methods has emerged as a promising way to obtain highly accurate energies and moderate scaling of computational cost. In this paper we significantly contribute towards this goal by introducing a novel deep-learning architecture that achieves 40-70\% lower energy error at 8x lower computational cost compared to previous approaches. Using our method we establish a new benchmark by calculating the most accurate variational ground state energies ever published for a number of different atoms and molecules. We systematically break down and measure our improvements, focusing in particular on the effect of increasing physical prior knowledge. We surprisingly find that increasing the prior knowledge given to the architecture can actually decrease accuracy.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DNN,FermiNet,ML-ESM,ML-QMBP,molecules,PauliNet,prediction of wavefunction,QMC,VMC},
file = {/Users/wasmer/Nextcloud/Zotero/Gerard et al_2022_Gold-standard solutions to the Schr-odinger equation using deep learning.pdf;/Users/wasmer/Zotero/storage/DWVRHXZW/2205.html}
}
@@ -5859,13 +6372,13 @@ Subject\_term: Quantum physics, Publishing, Peer review},
author = {Gerhorst, Christian-Roman and Neukirchen, Alexander and Klüppelberg, Daniel A. and Bihlmayer, Gustav and Betzinger, Markus and Michalicek, Gregor and Wortmann, Daniel and Blügel, Stefan},
date = {2023-09-26},
eprint = {2309.14799},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2309.14799},
url = {http://arxiv.org/abs/2309.14799},
urldate = {2023-10-04},
abstract = {Phonons are quantized vibrations of a crystal lattice that play a crucial role in understanding many properties of solids. Density functional theory (DFT) provides a state-of-the-art computational approach to lattice vibrations from first-principles. We present a successful software implementation for calculating phonons in the harmonic approximation, employing density-functional perturbation theory (DFPT) within the framework of the full-potential linearized augmented plane-wave (FLAPW) method as implemented in the electronic structure package FLEUR. The implementation, which involves the Sternheimer equation for the linear response of the wave function, charge density, and potential with respect to infinitesimal atomic displacements, as well as the setup of the dynamical matrix, is presented and the specifics due to the muffin-tin sphere centered LAPW basis-set and the all-electron nature are discussed. As a test, we calculate the phonon dispersion of several solids including an insulator, a semiconductor as well as several metals. The latter are comprised of magnetic, simple, and transition metals. The results are validated on the basis of phonon dispersions calculated using the finite displacement approach in conjunction with the FLEUR code and the phonopy package, as well as by some experimental results. An excellent agreement is obtained.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DFPT,DFT,DFT theory,FLEUR,FZJ,LAPW,perturbation theory,PGI,PGI-1/IAS-1,phonon,physics,quantum materials},
file = {/Users/wasmer/Nextcloud/Zotero/Gerhorst et al_2023_Phonons from Density-Functional Perturbation Theory using the All-Electron.pdf;/Users/wasmer/Zotero/storage/93USRLZ6/2309.html}
}
@@ -5900,7 +6413,7 @@ Subject\_term: Quantum physics, Publishing, Peer review},
url = {https://link.aps.org/doi/10.1103/PhysRevB.92.045131},
urldate = {2023-04-04},
abstract = {Based on an analysis of the short-range chemical environment of each atom in a system, standard machine-learning-based approaches to the construction of interatomic potentials aim at determining directly the central quantity, which is the total energy. This prevents, for instance, an accurate description of the energetics of systems in which long-range charge transfer or ionization is important. We propose therefore not to target directly with machine-learning methods the total energy but an intermediate physical quantity, namely, the charge density, which then in turn allows us to determine the total energy. By allowing the electronic charge to distribute itself in an optimal way over the system, we can describe not only neutral but also ionized systems with unprecedented accuracy. We demonstrate the power of our approach for both neutral and ionized NaCl clusters where charge redistribution plays a decisive role for the energetics. We are able to obtain chemical accuracy, i.e., errors of less than a millihartree per atom compared to the reference density functional results for a huge data set of configurations with large structural variety. The introduction of physically motivated quantities which are determined by the short-range atomic environment via a neural network also leads to an increased stability of the machine-learning process and transferability of the potential.},
- keywords = {\_tablet,AML,BigDFT,charge equilibration,charge transfer,electronegativity,electrostatic interaction,ionic systems,LDA,long-range interaction,ML,ML-DFT,MLP,NN,NNP,PES,prediction of electron density,prediction of electronegativity},
+ keywords = {AML,BigDFT,charge equilibration,charge transfer,electronegativity,electrostatic interaction,ionic systems,LDA,long-range interaction,ML,ML-DFT,MLP,NN,NNP,PES,prediction of electron density,prediction of electronegativity},
file = {/Users/wasmer/Nextcloud/Zotero/Ghasemi et al_2015_Interatomic potentials for ionic systems with density functional accuracy based.pdf;/Users/wasmer/Zotero/storage/HYUBBIPY/PhysRevB.92.html}
}
@@ -5978,17 +6491,36 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Ghosh, Kumar and Ghosh, Sumit},
date = {2022-07-26},
eprint = {2207.12837},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, physics:quant-ph},
doi = {10.48550/arXiv.2207.12837},
url = {http://arxiv.org/abs/2207.12837},
urldate = {2022-10-28},
abstract = {In this article we demonstrate the applications of classical and quantum machine learning in quantum transport and spintronics. With the help of a two terminal device with magnetic impurity we show how machine learning algorithms can predict the highly non-linear nature of conductance as well as the non-equilibrium spin response function for any random magnetic configuration. We finally describe the applicability of quantum machine learning which has the capability to handle a significantly large configuration space. Our approach is also applicable for molecular systems. These outcomes are crucial in predicting the behaviour of large scale systems where a quantum mechanical calculation is computationally challenging and therefore would play a crucial role in designing nano devices.},
- pubstate = {preprint},
- keywords = {\_tablet,FZJ,ML,PGI,PGI-1/IAS-1,QML,QSVM,quantum computing,quantum transport,random forest,rec-by-ghosh,spin dynamics,Spintronics,SVM,tight binding,transport properties},
+ pubstate = {prepublished},
+ keywords = {FZJ,ML,PGI,PGI-1/IAS-1,QML,QSVM,quantum computing,quantum transport,random forest,rec-by-ghosh,spin dynamics,Spintronics,SVM,tight binding,transport properties},
file = {/Users/wasmer/Nextcloud/Zotero/Ghosh_Ghosh_2022_Classical and quantum machine learning applications in spintronics.pdf;/Users/wasmer/Zotero/storage/FEUD8XZQ/2207.html}
}
+@article{ghoshExploringExoticConfigurations2023,
+ title = {Exploring Exotic Configurations with Anomalous Features with Deep Learning: {{Application}} of Classical and Quantum-Classical Hybrid Anomaly Detection},
+ shorttitle = {Exploring Exotic Configurations with Anomalous Features with Deep Learning},
+ author = {Ghosh, Kumar J. B. and Ghosh, Sumit},
+ date = {2023-10-09},
+ journaltitle = {Physical Review B},
+ shortjournal = {Phys. Rev. B},
+ volume = {108},
+ number = {16},
+ pages = {165408},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevB.108.165408},
+ url = {https://link.aps.org/doi/10.1103/PhysRevB.108.165408},
+ urldate = {2024-06-05},
+ abstract = {We present the application of classical and quantum-classical hybrid anomaly detection schemes to explore exotic configurations with anomalous features. We consider the Anderson model as a prototype, where we define two types of anomalies—a high conductance in the presence of strong impurity and a low conductance in the presence of weak impurity—as a function of random impurity distribution. Such anomalous outcome constitutes an imperceptible fraction of the data set and is not a part of the training process. These exotic configurations, which can be a source of rich new physics, usually remain elusive to conventional classification or regression methods and can be tracked only with a suitable anomaly detection scheme. We also present a systematic study of the performance of the classical and the quantum-classical hybrid anomaly detection method and show that the inclusion of a quantum circuit significantly enhances the performance of anomaly detection, which we quantify with suitable performance metrics. Our approach is quite generic in nature and can be used for any system that relies on a large number of parameters to find their new configurations, which can hold exotic new features.},
+ keywords = {condensed matter,FZJ,PGI,PGI-1/IAS-1,quantum machine learning,rec-by-ghosh},
+ file = {/Users/wasmer/Nextcloud/Zotero/Ghosh_Ghosh_2023_Exploring exotic configurations with anomalous features with deep learning.pdf;/Users/wasmer/Zotero/storage/T45ILR62/PhysRevB.108.html}
+}
+
@article{ghoshPerspectiveSpinOrbit2023,
title = {Perspective on Spin–Orbit Torque, Topology, and Reciprocal and Real-Space Spin Textures in Magnetic Materials and Heterostructures},
author = {Ghosh, Sumit and Rüßmann, Philipp and Mokrousov, Yuriy and Freimuth, Frank and Kosma, Adamantia},
@@ -6003,7 +6535,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
url = {https://doi.org/10.1063/5.0149849},
urldate = {2023-08-09},
abstract = {In this Perspective, we present some important aspects of two fundamental concepts of modern spintronics, namely, spin–orbit torque and topology. Although these two fields emerged separately in condensed matter physics, in spintronics they show a deep connection, which requires further theoretical and experimental investigation. The topological features can arise both from momentum space via the wave functions as well as from real space via complex magnetic configurations. These features manifest themselves as unique aspects of different equilibrium and non-equilibrium properties. Physical interactions of such a topological origin can open new possibilities for more efficient mechanisms for manipulating magnetic order with electrical currents, which, in turn, can lead to faster and more efficient spintronics devices.},
- keywords = {berry curvature,DFT,Dzyaloshinskii–Moriya interaction,FZJ,Hall AHE,Hall effect,Hall THE,impurity embedding,juKKR,Keldysh formalism,KKR,Kubo,Kubo-Bastin formalism,magnetic impurity,magnetic structure,magnetism,non-collinear,perspective,perspective-spintronics,PGI,PGI-1/IAS-1,physics,skyrmions,Spin-orbit effects,spin-orbit torque,spintronics,topological,topological insulator},
+ keywords = {\_tablet,berry curvature,DFT,Dzyaloshinskii–Moriya interaction,FZJ,Hall AHE,Hall effect,Hall THE,impurity embedding,juKKR,Keldysh formalism,KKR,Kubo,Kubo-Bastin formalism,magnetic impurity,magnetic structure,magnetism,non-collinear,perspective,perspective-spintronics,PGI,PGI-1/IAS-1,physics,skyrmions,Spin-orbit effects,spin-orbit torque,spintronics,topological,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Ghosh et al_2023_Perspective on spin–orbit torque, topology, and reciprocal and real-space spin.pdf;/Users/wasmer/Zotero/storage/UH6NDHMP/2896791.html}
}
@@ -6051,13 +6583,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Gilligan, Luke P. J. and Cobelli, Matteo and Taufour, Valentin and Sanvito, Stefano},
date = {2023-01-27},
eprint = {2301.11689},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2301.11689},
url = {http://arxiv.org/abs/2301.11689},
urldate = {2023-02-23},
abstract = {In recent times, transformer networks have achieved state-of-the-art performance in a wide range of natural language processing tasks. Here we present a workflow based on the fine-tuning of BERT models for different downstream tasks, which results in the automated extraction of structured information from unstructured natural language in scientific literature. Contrary to other methods for the automated extraction of structured compound-property relations from similar sources, our workflow does not rely on the definition of intricate grammar rules. Hence, it can be adapted to a new task without requiring extensive implementation efforts and knowledge. We test the data extraction performance by automatically generating a database of compounds and their associated Curie temperatures. This is compared with a manually curated database and one obtained with the state-of-the-art rule-based method. Finally, in order to demonstrate that the automatically extracted database can be used in a material-design workflow, we employ it to construct a machine-learning model predicting the Curie temperature based on a compound's chemical composition. This is quantitatively tested and compared with the best model constructed on manually-extracted data.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {\_tablet,BERT,data mining,database generation,literature analysis,LLM,materials},
file = {/Users/wasmer/Nextcloud/Zotero/Gilligan et al_2023_A rule-free workflow for the automated generation of databases from scientific.pdf;/Users/wasmer/Zotero/storage/W8WDMBDK/2301.html}
}
@@ -6067,13 +6599,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Gilmer, Justin and Schoenholz, Samuel S. and Riley, Patrick F. and Vinyals, Oriol and Dahl, George E.},
date = {2017-06-12},
eprint = {1704.01212},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.1704.01212},
url = {http://arxiv.org/abs/1704.01212},
urldate = {2022-10-03},
abstract = {Supervised learning on molecules has incredible potential to be useful in chemistry, drug discovery, and materials science. Luckily, several promising and closely related neural network models invariant to molecular symmetries have already been described in the literature. These models learn a message passing algorithm and aggregation procedure to compute a function of their entire input graph. At this point, the next step is to find a particularly effective variant of this general approach and apply it to chemical prediction benchmarks until we either solve them or reach the limits of the approach. In this paper, we reformulate existing models into a single common framework we call Message Passing Neural Networks (MPNNs) and explore additional novel variations within this framework. Using MPNNs we demonstrate state of the art results on an important molecular property prediction benchmark; these results are strong enough that we believe future work should focus on datasets with larger molecules or more accurate ground truth labels.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DeepMind,GCN,GNN,Google,ML,molecules,MPNN},
file = {/Users/wasmer/Nextcloud/Zotero/Gilmer et al_2017_Neural Message Passing for Quantum Chemistry.pdf;/Users/wasmer/Zotero/storage/A2EV2Y8T/1704.html}
}
@@ -6089,7 +6621,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
abstract = {Modern Condensed Matter Physics brings together the most important advances in the field of recent decades. It provides instructors teaching graduate-level condensed matter courses with a comprehensive and in-depth textbook that will prepare graduate students for research or further study as well as reading more advanced and specialized books and research literature in the field. This textbook covers the basics of crystalline solids as well as analogous optical lattices and photonic crystals, while discussing cutting-edge topics such as disordered systems, mesoscopic systems, many-body systems, quantum magnetism, Bose–Einstein condensates, quantum entanglement, and superconducting quantum bits. Students are provided with the appropriate mathematical background to understand the topological concepts that have been permeating the field, together with numerous physical examples ranging from the fractional quantum Hall effect to topological insulators, the toric code, and majorana fermions. Exercises, commentary boxes, and appendices afford guidance and feedback for beginners and experts alike.},
isbn = {978-1-316-48064-9},
langid = {english},
- keywords = {\_tablet,condensed matter,graduate,magnetism,superconductor,textbook,topological insulator},
+ keywords = {condensed matter,graduate,magnetism,superconductor,textbook,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Girvin_Yang_2019_Modern Condensed Matter Physics.pdf;/Users/wasmer/Zotero/storage/3FP65JQ3/F0A27AC5DEA8A40EA6EA5D727ED8B14E.html}
}
@@ -6117,7 +6649,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Glielmo, Aldo and Macocco, Iuri and Doimo, Diego and Carli, Matteo and Zeni, Claudio and Wild, Romina and family=Errico, given=Maria, prefix=d', useprefix=true and Rodriguez, Alex and Laio, Alessandro},
date = {2022-05-04},
eprint = {2205.03373},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
url = {http://arxiv.org/abs/2205.03373},
urldate = {2022-05-11},
@@ -6128,10 +6660,10 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
@unpublished{glielmoRankingInformationContent2021,
title = {Ranking the Information Content of Distance Measures},
- author = {Glielmo, Aldo and Zeni, Claudio and Cheng, Bingqing and Csanyi, Gabor and Laio, Alessandro},
+ author = {Glielmo, Aldo and Zeni, Claudio and Cheng, Bingqing and Csányi, Gábor and Laio, Alessandro},
date = {2021-04-30},
eprint = {2104.15079},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math, stat},
url = {http://arxiv.org/abs/2104.15079},
urldate = {2021-05-08},
@@ -6170,13 +6702,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Go, Dongwook and Lee, Hyun-Woo and Oppeneer, Peter M. and Blügel, Stefan and Mokrousov, Yuriy},
date = {2023-09-25},
eprint = {2309.13996},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2309.13996},
url = {http://arxiv.org/abs/2309.13996},
urldate = {2023-10-04},
abstract = {The position operator in a Bloch representation acquires a gauge correction in the momentum space on top of the canonical position, which is called the anomalous position. We show that the anomalous position is generally orbital-dependent and thus plays a crucial role in the description of the intrinsic orbital Hall effect in terms of Wannier basis. We demonstrate this from the first-principles calculation of orbital Hall conductivities of transition metals by Wannier interpolation. Our results show that consistent treatment of the velocity operator by adding the additional term originating from the anomalous position predicts the orbital Hall conductivities different from those obtained by considering only the group velocity. We find the difference is crucial in several metals. For example, we predict the negative sign of the orbital Hall conductivities for elements in the groups X and XI such as Cu, Ag, Au, and Pd, for which the previous studies predicted the positive sign. Our work suggests the importance of consistently describing the spatial dependence of basis functions by first-principles methods as it is fundamentally missing in the tight-binding approximation.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {DFT,FLEUR,FZJ,Hall effect,Hall OHE,Hall SHE,orbital angular momentum,PGI,PGI-1/IAS-1,physics,quantum materials,SOC,thin film,transition metals,Wannier},
file = {/Users/wasmer/Nextcloud/Zotero/Go et al_2023_First-principles calculation of orbital Hall effect by Wannier interpolation.pdf;/Users/wasmer/Zotero/storage/E39Y6NJ8/2309.html}
}
@@ -6205,13 +6737,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Gong, Chen and Maday, Yvon},
date = {2023-05-24},
eprint = {2305.14819},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math},
doi = {10.48550/arXiv.2305.14819},
url = {http://arxiv.org/abs/2305.14819},
urldate = {2023-07-10},
abstract = {Molecular representation learning (MRL) has long been crucial in the fields of drug discovery and materials science, and it has made significant progress due to the development of natural language processing (NLP) and graph neural networks (GNNs). NLP treats the molecules as one dimensional sequential tokens while GNNs treat them as two dimensional topology graphs. Based on different message passing algorithms, GNNs have various performance on detecting chemical environments and predicting molecular properties. Herein, we propose Directed Graph Attention Networks (D-GATs): the expressive GNNs with directed bonds. The key to the success of our strategy is to treat the molecular graph as directed graph and update the bond states and atom states by scaled dot-product attention mechanism. This allows the model to better capture the sub-structure of molecular graph, i.e., functional groups. Compared to other GNNs or Message Passing Neural Networks (MPNNs), D-GATs outperform the state-of-the-art on 13 out of 15 important molecular property prediction benchmarks.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,Computer Science - Information Theory},
file = {/Users/wasmer/Nextcloud/Zotero/Gong_Maday_2023_Directed Message Passing Based on Attention for Prediction of Molecular.pdf;/Users/wasmer/Zotero/storage/YKSQR6HZ/2305.html}
}
@@ -6233,10 +6765,28 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
abstract = {The combination of deep learning and ab initio calculation has shown great promise in revolutionizing future scientific research, but how to design neural network models incorporating a priori knowledge and symmetry requirements is a key challenging subject. Here we propose an E(3)-equivariant deep-learning framework to represent density functional theory (DFT) Hamiltonian as a function of material structure, which can naturally preserve the Euclidean symmetry even in the presence of spin–orbit coupling. Our DeepH-E3 method enables efficient electronic structure calculation at ab initio accuracy by learning from DFT data of small-sized structures, making the routine study of large-scale supercells ({$>$}104 atoms) feasible. The method can reach sub-meV prediction accuracy at high training efficiency, showing state-of-the-art performance in our experiments. The work is not only of general significance to deep-learning method development but also creates opportunities for materials research, such as building a Moiré-twisted material database.},
issue = {1},
langid = {english},
- keywords = {\_tablet,AML,bismuth selenide,bismuth telluride,DeepH,E(3),e3nn,ENN,equivariant,library,materials,ML,ML-DFT,ML-ESM,PAW,PBE,prediction of Hamiltonian matrix,PyTorch,SOC,spin-dependent,topological insulator,twisted bilayer graphene,VASP,vdW materials,with-code,with-data},
+ keywords = {AML,bismuth selenide,bismuth telluride,DeepH,E(3),e3nn,ENN,equivariant,library,materials,ML,ML-DFT,ML-ESM,PAW,PBE,prediction of Hamiltonian matrix,PyTorch,SOC,spin-dependent,topological insulator,twisted bilayer graphene,VASP,vdW materials,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Gong et al_2023_General framework for E(3)-equivariant neural network representation of density.pdf;/Users/wasmer/Nextcloud/Zotero/Gong et al_2023_General framework for E(3)-equivariant neural network representation of density2.pdf;/Users/wasmer/Nextcloud/Zotero/Gong et al_2023_General framework for E(3)-equivariant neural network representation of density3.pdf;/Users/wasmer/Nextcloud/Zotero/Gong et al_2023_General framework for E(3)-equivariant neural network representation of density4.pdf}
}
+@article{gongPredictingChargeDensity2019,
+ title = {Predicting Charge Density Distribution of Materials Using a Local-Environment-Based Graph Convolutional Network},
+ author = {Gong, Sheng and Xie, Tian and Zhu, Taishan and Wang, Shuo and Fadel, Eric R. and Li, Yawei and Grossman, Jeffrey C.},
+ date = {2019-11-07},
+ journaltitle = {Physical Review B},
+ shortjournal = {Phys. Rev. B},
+ volume = {100},
+ number = {18},
+ pages = {184103},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevB.100.184103},
+ url = {https://link.aps.org/doi/10.1103/PhysRevB.100.184103},
+ urldate = {2024-05-27},
+ abstract = {The electron charge density distribution of materials is one of the key quantities in computational materials science as theoretically it determines the ground state energy and practically it is used in many materials analyses. However, the scaling of density functional theory calculations with number of atoms limits the usage of charge-density-based calculations and analyses. Here we introduce a machine-learning scheme with local-environment-based graphs and graph convolutional neural networks to predict charge density on grid points from the crystal structure. We show the accuracy of this scheme through a comparison of predicted charge densities as well as properties derived from the charge density, and that the scaling is ð‘‚(ð‘). More importantly, the transferability is shown to be high with respect to different compositions and structures, which results from the explicit encoding of geometry.},
+ keywords = {AML,CGCNN,crystal graph,GNN,ML,ML-Density,ML-DFT,ML-ESM,prediction of electron density},
+ file = {/Users/wasmer/Nextcloud/Zotero/Gong et al_2019_Predicting charge density distribution of materials using a.pdf;/Users/wasmer/Zotero/storage/DCU9YN73/PhysRevB.100.html}
+}
+
@book{gonisMultipleScatteringSolids2000,
title = {Multiple {{Scattering}} in {{Solids}}},
author = {Gonis, Antonios and Butler, William H.},
@@ -6260,7 +6810,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Goodall, Rhys E. A. and Parackal, Abhijith S. and Faber, Felix A. and Armiento, Rickard and Lee, Alpha A.},
date = {2022-03-15},
eprint = {2106.11132},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2106.11132},
urldate = {2022-05-09},
@@ -6274,13 +6824,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Goodall, Rhys E. A. and Parackal, Abhijith S. and Faber, Felix A. and Armiento, Rickard and Lee, Alpha A.},
date = {2022-03-15},
eprint = {2106.11132},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2106.11132},
url = {http://arxiv.org/abs/2106.11132},
urldate = {2022-10-03},
abstract = {A fundamental challenge in materials science pertains to elucidating the relationship between stoichiometry, stability, structure, and property. Recent advances have shown that machine learning can be used to learn such relationships, allowing the stability and functional properties of materials to be accurately predicted. However, most of these approaches use atomic coordinates as input and are thus bottle-necked by crystal structure identification when investigating novel materials. Our approach solves this bottleneck by coarse-graining the infinite search space of atomic coordinates into a combinatorially enumerable search space. The key idea is to use Wyckoff representations -- coordinate-free sets of symmetry-related positions in a crystal -- as the input to a machine learning model. Our model demonstrates exceptionally high precision in discovering new theoretically stable materials, identifying 1,569 materials that lie below the known convex hull of previously calculated materials from just 5,675 ab-initio calculations. Our approach opens up fundamental advances in computational materials discovery.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {CGCNN,Condensed Matter - Materials Science,crystal structure,crystal symmetry,GNN,MPNN,original publication,Physics - Computational Physics,regression,Wren,Wyckoff positions,Wyckoff representation},
file = {/Users/wasmer/Nextcloud/Zotero/Goodall et al_2022_Rapid Discovery of Stable Materials by Coordinate-free Coarse Graining2.pdf;/Users/wasmer/Zotero/storage/8I7WCWRJ/2106.html}
}
@@ -6314,13 +6864,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Go, Dongwook and Ando, Kazuya and Pezo, Armando and Blügel, Stefan and Manchon, Aurélien and Mokrousov, Yuriy},
date = {2023-09-26},
eprint = {2309.14817},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2309.14817},
url = {http://arxiv.org/abs/2309.14817},
urldate = {2023-10-04},
abstract = {We show that dynamics of the magnetization in ferromagnets can pump the orbital angular momentum, which we denote by orbital pumping. This is the reciprocal phenomenon to the orbital torque that induces magnetization dynamics by the orbital angular momentum in non-equilibrium. The orbital pumping is analogous to the spin pumping established in spintronics but requires the spin-orbit coupling for the orbital angular momentum to interact with the magnetization. We develop a formalism that describes the generation of the orbital angular momentum by magnetization dynamics within the adiabatic perturbation theory. Based on this, we perform first-principles calculation of the orbital pumping in prototypical \$3d\$ ferromagnets, Fe, Co, and Ni. The results show that the ratio between the orbital pumping and the spin pumping ranges from 5 to 15 percents, being smallest in Fe and largest in Ni. This implies that ferromagnetic Ni is a good candidate for measuring the orbital pumping. Implications of our results on experiments are also discussed.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Ferromagnetism,FZJ,magnetization,orbital angular momentum,PGI,PGI-1/IAS-1,physics,pumping,quantum materials,SOC,transition metals},
file = {/Users/wasmer/Nextcloud/Zotero/Go et al_2023_Orbital Pumping by Magnetization Dynamics in Ferromagnets.pdf;/Users/wasmer/Zotero/storage/EHVUDCNY/2309.html}
}
@@ -6344,7 +6894,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Goscinski, Alexander and Musil, Félix and Pozdnyakov, Sergey and Ceriotti, Michele},
date = {2021-05-18},
eprint = {2105.08717},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
url = {http://arxiv.org/abs/2105.08717},
urldate = {2021-05-30},
@@ -6382,8 +6932,8 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
url = {https://open-research-europe.ec.europa.eu/articles/3-81/v2},
urldate = {2023-10-01},
abstract = {Easy-to-use libraries such as scikit-learn have accelerated the adoption and application of machine learning (ML) workflows and data-driven methods. While many of the algorithms implemented in these libraries originated in specific scientific fields, they have gained in popularity in part because of their generalisability across multiple domains. Over the past two decades, researchers in the chemical and materials science community have put forward general-purpose machine learning methods. The deployment of these methods into workflows of other domains, however, is often burdensome due to the entanglement with domain-specific functionalities. We present the python library scikit-matter that targets domain-agnostic implementations of methods developed in the computational chemical and materials science community, following the scikit-learn API and coding guidelines to promote usability and interoperability with existing workflows.},
- pubstate = {preprint},
- keywords = {\_tablet,AML,convex hull,CUR decomposition,descriptors,dimensionality reduction,error estimate,farthest point sampling,feature reconstuction measure,feature selection,kernel methods,library,ML,PCovR,regression,sample selection,scikit-learn,SOAP,unsupervised learning,with-code,with-data},
+ pubstate = {prepublished},
+ keywords = {AML,convex hull,CUR decomposition,descriptors,dimensionality reduction,error estimate,farthest point sampling,feature reconstuction measure,feature selection,kernel methods,library,ML,PCovR,regression,sample selection,scikit-learn,SOAP,unsupervised learning,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Goscinski et al_2023_A Suite of Generalisable Machine Learning Methods Born out of Chemistry and.pdf;/Users/wasmer/Zotero/storage/IAJ8UA3M/v2.html}
}
@@ -6392,33 +6942,17 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Grassano, Davide and Marzari, Nicola and Campi, Davide},
date = {2023-08-04},
eprint = {2308.01663},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2308.01663},
url = {http://arxiv.org/abs/2308.01663},
urldate = {2023-10-08},
abstract = {Topological Weyl semimetals represent a novel class of non-trivial materials, where band crossings with linear dispersions take place at generic momenta across reciprocal space. These crossings give rise to low-energy properties akin to those of Weyl fermions, and are responsible for several exotic phenomena. Up to this day, only a handful of Weyl semimetals have been discovered, and the search for new ones remains a very active area. The main challenge on the computational side arises from the fact that many of the tools used to identify the topological class of a material do not provide a complete picture in the case of Weyl semimetals. In this work, we propose an alternative and inexpensive, criterion to screen for possible Weyl fermions, based on the analysis of the band structure along high-symmetry directions in the absence of spin-orbit coupling. We test the method by running a high-throughput screening on a set of 5455 inorganic bulk materials and identify 49 possible candidates for topological properties. A further analysis, carried out by identifying and characterizing the crossings in the Brillouin zone, shows us that 3 of these candidates are Weyl semimetals. Interestingly, while these 3 materials underwent other high-throughput screenings, none had revealed their topological behavior before.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AiiDA,DFT,Hall effect,Hall QAHE,High-throughput,HTC,Materials Cloud,materials screening,PBE,physics,Quantum ESPRESSO,quantum materials,SOC,topological,Topological matter,TRS,Weyl semimetal,with-code,with-data,workflows},
file = {/Users/wasmer/Nextcloud/Zotero/Grassano et al_2023_High-throughput screening of Weyl semimetals.pdf;/Users/wasmer/Zotero/storage/QWHL9R54/2308.html}
}
-@online{grisafiElectronicstructurePropertiesAtomcentered2022,
- title = {Electronic-Structure Properties from Atom-Centered Predictions of the Electron Density},
- author = {Grisafi, Andrea and Lewis, Alan M. and Rossi, Mariana and Ceriotti, Michele},
- date = {2022-06-28},
- eprint = {2206.14087},
- eprinttype = {arxiv},
- eprintclass = {cond-mat, physics:physics, stat},
- doi = {10.48550/arXiv.2206.14087},
- url = {http://arxiv.org/abs/2206.14087},
- urldate = {2022-07-02},
- abstract = {The electron density of a molecule or material has recently received major attention as a target quantity of machine-learning models. A natural choice to construct a model that yields transferable and linear-scaling predictions is to represent the scalar field using a multi-centered atomic basis analogous to that routinely used in density fitting approximations. However, the non-orthogonality of the basis poses challenges for the learning exercise, as it requires accounting for all the atomic density components at once. We devise a gradient-based approach to directly minimize the loss function of the regression problem in an optimized and highly sparse feature space. In so doing, we overcome the limitations associated with adopting an atom-centered model to learn the electron density over arbitrarily complex datasets, obtaining extremely accurate predictions. The enhanced framework is tested on 32-molecule periodic cells of liquid water, presenting enough complexity to require an optimal balance between accuracy and computational efficiency. We show that starting from the predicted density a single Kohn-Sham diagonalization step can be performed to access total energy components that carry an error of just 0.1 meV/atom with respect to the reference density functional calculations. Finally, we test our method on the highly heterogeneous QM9 benchmark dataset, showing that a small fraction of the training data is enough to derive ground-state total energies within chemical accuracy.},
- pubstate = {preprint},
- keywords = {\_tablet,DFT,dimensionality reduction,equivariant,lambda-SOAP,ML-DFT,ML-ESM,molecules,molecules \& solids,prediction of electron density,QM9,RKHS,SA-GPR,SALTED,SOAP,solids},
- file = {/Users/wasmer/Nextcloud/Zotero/Grisafi et al_2022_Electronic-structure properties from atom-centered predictions of the electron.pdf;/Users/wasmer/Zotero/storage/QPHBS33I/2206.html}
-}
-
@article{grisafiElectronicStructurePropertiesAtomCentered2022,
title = {Electronic-{{Structure Properties}} from {{Atom-Centered Predictions}} of the {{Electron Density}}},
author = {Grisafi, Andrea and Lewis, Alan M. and Rossi, Mariana and Ceriotti, Michele},
@@ -6435,14 +6969,30 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
file = {/Users/wasmer/Nextcloud/Zotero/Grisafi et al_2022_Electronic-Structure Properties from Atom-Centered Predictions of the Electron.pdf;/Users/wasmer/Zotero/storage/29HAHUDS/acs.jctc.html}
}
-@article{grisafiIncorporatingLongrangePhysics2019,
- title = {Incorporating Long-Range Physics in Atomic-Scale Machine Learning},
- author = {Grisafi, Andrea and Ceriotti, Michele},
- date = {2019-11-28},
- journaltitle = {The Journal of Chemical Physics},
- shortjournal = {J. Chem. Phys.},
- volume = {151},
- number = {20},
+@online{grisafiElectronicstructurePropertiesAtomcentered2022a,
+ title = {Electronic-Structure Properties from Atom-Centered Predictions of the Electron Density},
+ author = {Grisafi, Andrea and Lewis, Alan M. and Rossi, Mariana and Ceriotti, Michele},
+ date = {2022-06-28},
+ eprint = {2206.14087},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat, physics:physics, stat},
+ doi = {10.48550/arXiv.2206.14087},
+ url = {http://arxiv.org/abs/2206.14087},
+ urldate = {2022-07-02},
+ abstract = {The electron density of a molecule or material has recently received major attention as a target quantity of machine-learning models. A natural choice to construct a model that yields transferable and linear-scaling predictions is to represent the scalar field using a multi-centered atomic basis analogous to that routinely used in density fitting approximations. However, the non-orthogonality of the basis poses challenges for the learning exercise, as it requires accounting for all the atomic density components at once. We devise a gradient-based approach to directly minimize the loss function of the regression problem in an optimized and highly sparse feature space. In so doing, we overcome the limitations associated with adopting an atom-centered model to learn the electron density over arbitrarily complex datasets, obtaining extremely accurate predictions. The enhanced framework is tested on 32-molecule periodic cells of liquid water, presenting enough complexity to require an optimal balance between accuracy and computational efficiency. We show that starting from the predicted density a single Kohn-Sham diagonalization step can be performed to access total energy components that carry an error of just 0.1 meV/atom with respect to the reference density functional calculations. Finally, we test our method on the highly heterogeneous QM9 benchmark dataset, showing that a small fraction of the training data is enough to derive ground-state total energies within chemical accuracy.},
+ pubstate = {prepublished},
+ keywords = {DFT,dimensionality reduction,equivariant,lambda-SOAP,ML-DFT,ML-ESM,molecules,molecules \& solids,prediction of electron density,QM9,RKHS,SA-GPR,SALTED,SOAP,solids},
+ file = {/Users/wasmer/Nextcloud/Zotero/Grisafi et al_2022_Electronic-structure properties from atom-centered predictions of the electron.pdf;/Users/wasmer/Zotero/storage/QPHBS33I/2206.html}
+}
+
+@article{grisafiIncorporatingLongrangePhysics2019,
+ title = {Incorporating Long-Range Physics in Atomic-Scale Machine Learning},
+ author = {Grisafi, Andrea and Ceriotti, Michele},
+ date = {2019-11-28},
+ journaltitle = {The Journal of Chemical Physics},
+ shortjournal = {J. Chem. Phys.},
+ volume = {151},
+ number = {20},
pages = {204105},
publisher = {American Institute of Physics},
issn = {0021-9606},
@@ -6498,13 +7048,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Grisafi, Andrea and Bussy, Augustin and Vuilleumier, Rodolphe},
date = {2023-04-18},
eprint = {2304.08966},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2304.08966},
url = {http://arxiv.org/abs/2304.08966},
urldate = {2023-05-26},
abstract = {The computational study of energy storage and conversion processes call for simulation techniques that can reproduce the electronic response of metal electrodes under electric fields. Despite recent advancements in machine-learning methods applied to electronic-structure properties, predicting the non-local behaviour of the charge density in electronic conductors remains a major open challenge. We combine long-range and equivariant kernel methods to predict the Kohn-Sham electron density of metal electrodes decomposed on an atom-centered basis. By taking slabs of gold as an example, we show that including long-range correlations into the learning model is essential to accurately reproduce the charge density and potential in bare electrodes of increasing size. A finite-field extension of the method is then introduced, which allows us to predict the charge transfer and the electrostatic potential drop induced by the application of an external electric field. Finally, we demonstrate the capability of the method to extrapolate the non-local electronic polarization generated by the interaction with an ionic species for electrodes of arbitrary thickness. Our study represents an important step forward in the accurate simulation of energy materials that include metallic interfaces.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,DFT,electrochemistry,electrode,equivariant,GPR,kernel methods,LODE,long-range interaction,ML,ML-DFT,ML-ESM,prediction of electron density,SALTED,SOAP},
file = {/Users/wasmer/Nextcloud/Zotero/Grisafi et al_2023_Predicting the Charge Density Response in Metal Electrodes.pdf;/Users/wasmer/Zotero/storage/ZDZULVEX/2304.html}
}
@@ -6542,7 +7092,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
url = {https://doi.org/10.1021/acscentsci.8b00551},
urldate = {2021-10-19},
abstract = {The electronic charge density plays a central role in determining the behavior of matter at the atomic scale, but its computational evaluation requires demanding electronic-structure calculations. We introduce an atom-centered, symmetry-adapted framework to machine-learn the valence charge density based on a small number of reference calculations. The model is highly transferable, meaning it can be trained on electronic-structure data of small molecules and used to predict the charge density of larger compounds with low, linear-scaling cost. Applications are shown for various hydrocarbon molecules of increasing complexity and flexibility, and demonstrate the accuracy of the model when predicting the density on octane and octatetraene after training exclusively on butane and butadiene. This transferable, data-driven model can be used to interpret experiments, accelerate electronic structure calculations, and compute electrostatic interactions in molecules and condensed-phase systems.},
- keywords = {\_tablet,DFT,GPR,lambda-SOAP,library,ML,ML-DFT,ML-ESM,models,molecules,prediction of electron density,SA-GPR,SALTED,SOAP,with-code},
+ keywords = {DFT,GPR,lambda-SOAP,library,ML,ML-DFT,ML-ESM,models,molecules,prediction of electron density,SA-GPR,SALTED,SOAP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Grisafi et al_2019_Transferable Machine-Learning Model of the Electron Density.pdf;/Users/wasmer/Zotero/storage/HMBCGARZ/acscentsci.html}
}
@@ -6566,17 +7116,37 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Gruver, Nate and Sriram, Anuroop and Madotto, Andrea and Wilson, Andrew Gordon and Zitnick, C. Lawrence and Ulissi, Zachary},
date = {2024-02-06},
eprint = {2402.04379},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2402.04379},
url = {http://arxiv.org/abs/2402.04379},
urldate = {2024-05-07},
abstract = {We propose fine-tuning large language models for generation of stable materials. While unorthodox, fine-tuning large language models on text-encoded atomistic data is simple to implement yet reliable, with around 90\% of sampled structures obeying physical constraints on atom positions and charges. Using energy above hull calculations from both learned ML potentials and gold-standard DFT calculations, we show that our strongest model (fine-tuned LLaMA-2 70B) can generate materials predicted to be metastable at about twice the rate (49\% vs 28\%) of CDVAE, a competing diffusion model. Because of text prompting's inherent flexibility, our models can simultaneously be used for unconditional generation of stable material, infilling of partial structures and text-conditional generation. Finally, we show that language models' ability to capture key symmetries of crystal structures improves with model scale, suggesting that the biases of pretrained LLMs are surprisingly well-suited for atomistic data.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,crystal structure prediction,fine-tuning,language models,LLM,materials discovery,Meta Research,ML,pretrained models,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Gruver et al_2024_Fine-Tuned Language Models Generate Stable Inorganic Materials as Text.pdf;/Users/wasmer/Zotero/storage/J53W542H/2402.html}
}
+@article{gubaevPerformanceTwoComplementary2023,
+ title = {Performance of Two Complementary Machine-Learned Potentials in Modelling Chemically Complex Systems},
+ author = {Gubaev, Konstantin and Zaverkin, Viktor and Srinivasan, Prashanth and Duff, Andrew Ian and Kästner, Johannes and Grabowski, Blazej},
+ date = {2023-07-25},
+ journaltitle = {npj Computational Materials},
+ shortjournal = {npj Comput Mater},
+ volume = {9},
+ number = {1},
+ pages = {1--15},
+ publisher = {Nature Publishing Group},
+ issn = {2057-3960},
+ doi = {10.1038/s41524-023-01073-w},
+ url = {https://www.nature.com/articles/s41524-023-01073-w},
+ urldate = {2024-06-15},
+ abstract = {Chemically complex multicomponent alloys possess exceptional properties derived from an inexhaustible compositional space. The complexity however makes interatomic potential development challenging. We explore two complementary machine-learned potentials—the moment tensor potential (MTP) and the Gaussian moment neural network (GM-NN)—in simultaneously describing configurational and vibrational degrees of freedom in the Ta-V-Cr-W alloy family. Both models are equally accurate with excellent performance evaluated against density-functional-theory. They achieve root-mean-square-errors (RMSEs) in energies of less than a few meV/atom across 0 K ordered and high-temperature disordered configurations included in the training. Even for compositions not in training, relative energy RMSEs at high temperatures are within a few meV/atom. High-temperature molecular dynamics forces have similarly small RMSEs of about 0.15 eV/Å for the disordered quaternary included in, and ternaries not part of training. MTPs achieve faster convergence with training size; GM-NNs are faster in execution. Active learning is partially beneficial and should be complemented with conventional human-based training set generation.},
+ langid = {english},
+ keywords = {/unread,active learning,alloys,AML,descriptors,GM-NN,GTO basis,HEA,library,ML,MLP,model comparison,MTP,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Gubaev et al_2023_Performance of two complementary machine-learned potentials in modelling.pdf}
+}
+
@book{gubanovMagnetismElectronicStructure1992,
title = {Magnetism and the {{Electronic Structure}} of {{Crystals}}},
author = {Gubanov, V. A. and Liechtenstein, A. I. and Postnikov, A. V.},
@@ -6588,7 +7158,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
urldate = {2022-06-18},
isbn = {978-3-642-84413-3},
langid = {english},
- keywords = {\_tablet,condensed matter,defects,DFT,magnetism},
+ keywords = {condensed matter,defects,DFT,magnetism},
file = {/Users/wasmer/Nextcloud/Zotero/Magnetism and the Electronic Structure of Crystals.pdf;/Users/wasmer/Zotero/storage/QVJRNHRA/978-3-642-84411-9.html}
}
@@ -6628,13 +7198,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Gutmann, Michael U.},
date = {2022-06-27},
eprint = {2206.13446},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.2206.13446},
url = {http://arxiv.org/abs/2206.13446},
urldate = {2022-06-29},
abstract = {This is a collection of (mostly) pen-and-paper exercises in machine learning. The exercises are on the following topics: linear algebra, optimisation, directed graphical models, undirected graphical models, expressive power of graphical models, factor graphs and message passing, inference for hidden Markov models, model-based learning (including ICA and unnormalised models), sampling and Monte-Carlo integration, and variational inference.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {course material,exercises,General ML,graphical model,hidden Markov model,MC integration,ML,sampling,teaching},
file = {/Users/wasmer/Nextcloud/Zotero/Gutmann_2022_Pen and Paper Exercises in Machine Learning.pdf;/Users/wasmer/Zotero/storage/KMSFX6RY/2206.html}
}
@@ -6731,6 +7301,19 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
file = {/Users/wasmer/Nextcloud/Zotero/Hafner_2008_Ab-initio simulations of materials using VASP.pdf;/Users/wasmer/Zotero/storage/UCE26HNQ/jcc.html}
}
+@article{hajkowiczArtificialIntelligenceScience2022,
+ title = {Artificial {{Intelligence}} for {{Science}} - {{Adoption Trends}} and {{Future Development Pathways}}},
+ author = {Hajkowicz, Stefan and Naughtin, Claire and Sanderson, Conrad and Schleiger, Emma and Karimi, Sarvnaz and Bratanova, Alexandra and Bednarz, Tomasz},
+ date = {2022},
+ publisher = {[object Object]},
+ doi = {10.13140/RG.2.2.20389.88800},
+ url = {https://www.csiro.au/en/research/technology-space/ai/Artificial-Intelligence-for-Science-report},
+ urldate = {2024-05-16},
+ langid = {english},
+ keywords = {/unread,AI,AI4Science,chemistry,CSIRO,DeepMind,materials,Microsoft Research,ML,physical sciences,physics,report,review-of-AI4science,science},
+ file = {/Users/wasmer/Nextcloud/Zotero/Hajkowicz et al_2022_Artificial Intelligence for Science - Adoption Trends and Future Development.pdf}
+}
+
@inproceedings{hammermeshGroupTheoryIts1963,
title = {\emph{Group }{{\emph{Theory}}}\emph{ and }{{\emph{Its Application}}}\emph{ to }{{\emph{Physical Problems}}}},
booktitle = {Physics {{Today}}},
@@ -6772,13 +7355,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Hanson, Mark A. and Barreiro, Pablo Gómez and Crosetto, Paolo and Brockington, Dan},
date = {2023-09-27},
eprint = {2309.15884},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2309.15884},
url = {http://arxiv.org/abs/2309.15884},
urldate = {2023-10-02},
abstract = {Scientists are increasingly overwhelmed by the volume of articles being published. Total articles indexed in Scopus and Web of Science have grown exponentially in recent years; in 2022 the article total was 47\% higher than in 2016, which has outpaced the limited growth, if any, in the number of practising scientists. Thus, publication workload per scientist (writing, reviewing, editing) has increased dramatically. We define this problem as the strain on scientific publishing. To analyse this strain, we present five data-driven metrics showing publisher growth, processing times, and citation behaviours. We draw these data from web scrapes, requests for data from publishers, and material that is freely available through publisher websites. Our findings are based on millions of papers produced by leading academic publishers. We find specific groups have disproportionately grown in their articles published per year, contributing to this strain. Some publishers enabled this growth by adopting a strategy of hosting special issues, which publish articles with reduced turnaround times. Given pressures on researchers to publish or perish to be competitive for funding applications, this strain was likely amplified by these offers to publish more articles. We also observed widespread year-over-year inflation of journal impact factors coinciding with this strain, which risks confusing quality signals. Such exponential growth cannot be sustained. The metrics we define here should enable this evolving conversation to reach actionable solutions to address the strain on scientific publishing.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,academia,criticism,literature analysis,publishing,scientific journals,working in science},
file = {/Users/wasmer/Nextcloud/Zotero/Hanson et al_2023_The strain on scientific publishing.pdf;/Users/wasmer/Zotero/storage/YH795P59/2309.html}
}
@@ -6836,7 +7419,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
url = {https://link.aps.org/doi/10.1103/RevModPhys.82.3045},
urldate = {2023-06-15},
abstract = {Topological insulators are electronic materials that have a bulk band gap like an ordinary insulator but have protected conducting states on their edge or surface. These states are possible due to the combination of spin-orbit interactions and time-reversal symmetry. The two-dimensional (2D) topological insulator is a quantum spin Hall insulator, which is a close cousin of the integer quantum Hall state. A three-dimensional (3D) topological insulator supports novel spin-polarized 2D Dirac fermions on its surface. In this Colloquium the theoretical foundation for topological insulators and superconductors is reviewed and recent experiments are described in which the signatures of topological insulators have been observed. Transport experiments on HgTe∕CdTe quantum wells are described that demonstrate the existence of the edge states predicted for the quantum spin Hall insulator. Experiments on Bi1−xSbx, Bi2Se3, Bi2Te3, and Sb2Te3 are then discussed that establish these materials as 3D topological insulators and directly probe the topology of their surface states. Exotic states are described that can occur at the surface of a 3D topological insulator due to an induced energy gap. A magnetic gap leads to a novel quantum Hall state that gives rise to a topological magnetoelectric effect. A superconducting energy gap leads to a state that supports Majorana fermions and may provide a new venue for realizing proposals for topological quantum computation. Prospects for observing these exotic states are also discussed, as well as other potential device applications of topological insulators.},
- keywords = {\_tablet,2D,3D,BdG,Bi2Te3,bismuth selenide,bismuth telluride,Chern number,colloquium,graphene,groundbreaking,Hall effect,Hall QAHE,Hall QSHE,Majorana,MZM,superconductor,TKNN,topological,topological insulator,TRS},
+ keywords = {2D,3D,BdG,Bi2Te3,bismuth selenide,bismuth telluride,Chern number,colloquium,for introductions,graphene,groundbreaking,Hall effect,Hall QAHE,Hall QSHE,Majorana,MZM,superconductor,TKNN,topological,topological insulator,TRS},
file = {/Users/wasmer/Nextcloud/Zotero/Hasan_Kane_2010_Colloquium.pdf;/Users/wasmer/Zotero/storage/RXPD79NW/RevModPhys.82.html}
}
@@ -6853,26 +7436,62 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
url = {https://royalsocietypublishing.org/doi/10.1098/rsta.2013.0270},
urldate = {2022-05-18},
abstract = {Density functional theory (DFT) has been used in many fields of the physical sciences, but none so successfully as in the solid state. From its origins in condensed matter physics, it has expanded into materials science, high-pressure physics and mineralogy, solid-state chemistry and more, powering entire computational subdisciplines. Modern DFT simulation codes can calculate a vast range of structural, chemical, optical, spectroscopic, elastic, vibrational and thermodynamic phenomena. The ability to predict structure–property relationships has revolutionized experimental fields, such as vibrational and solid-state NMR spectroscopy, where it is the primary method to analyse and interpret experimental spectra. In semiconductor physics, great progress has been made in the electronic structure of bulk and defect states despite the severe challenges presented by the description of excited states. Studies are no longer restricted to known crystallographic structures. DFT is increasingly used as an exploratory tool for materials discovery and computational experiments, culminating in ex nihilo crystal structure prediction, which addresses the long-standing difficult problem of how to predict crystal structure polymorphs from nothing but a specified chemical composition. We present an overview of the capabilities of solid-state DFT simulations in all of these topics, illustrated with recent examples using the CASTEP computer program.},
- keywords = {\_tablet,condensed matter,DFT,review},
+ keywords = {condensed matter,DFT,review},
file = {/Users/wasmer/Nextcloud/Zotero/Hasnip et al_2014_Density functional theory in the solid state.pdf}
}
+@online{hayesSimulating500Million2024,
+ title = {Simulating 500 Million Years of Evolution with a Language Model},
+ author = {Hayes, Thomas and Rao, Roshan and Akin, Halil and Sofroniew, Nicholas J. and Oktay, Deniz and Lin, Zeming and Verkuil, Robert and Tran, Vincent Q. and Deaton, Jonathan and Wiggert, Marius and Badkundri, Rohil and Shafkat, Irhum and Gong, Jun and Derry, Alexander and Molina, Raul S. and Thomas, Neil and Khan, Yousuf and Mishra, Chetan and Kim, Carolyn and Bartie, Liam J. and Nemeth, Matthew and Hsu, Patrick D. and Sercu, Tom and Candido, Salvatore and Rives, Alexander},
+ date = {2024-07-02},
+ eprinttype = {bioRxiv},
+ eprintclass = {New Results},
+ pages = {2024.07.01.600583},
+ doi = {10.1101/2024.07.01.600583},
+ url = {https://www.biorxiv.org/content/10.1101/2024.07.01.600583v1},
+ urldate = {2024-08-03},
+ abstract = {More than three billion years of evolution have produced an image of biology encoded into the space of natural proteins. Here we show that language models trained on tokens generated by evolution can act as evolutionary simulators to generate functional proteins that are far away from known proteins. We present ESM3, a frontier multimodal generative language model that reasons over the sequence, structure, and function of proteins. ESM3 can follow complex prompts combining its modalities and is highly responsive to biological alignment. We have prompted ESM3 to generate fluorescent proteins with a chain of thought. Among the generations that we synthesized, we found a bright fluorescent protein at far distance (58\% identity) from known fluorescent proteins. Similarly distant natural fluorescent proteins are separated by over five hundred million years of evolution.},
+ langid = {english},
+ pubstate = {prepublished},
+ keywords = {/unread,AI4Science,biomolecules,generative models,groundbreaking,language models,LLM,Protein structure predictions,structure prediction,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Hayes et al_2024_Simulating 500 million years of evolution with a language model.pdf}
+}
+
@online{hazraPredictingOneParticleDensity2024,
title = {Predicting {{The One-Particle Density Matrix With Machine Learning}}},
author = {Hazra, S. and Patil, U. and Sanvito, S.},
date = {2024-01-12},
eprint = {2401.06533},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2401.06533},
url = {http://arxiv.org/abs/2401.06533},
urldate = {2024-03-08},
abstract = {Two of the most widely used electronic structure theory methods, namely Hartree-Fock and Kohn-Sham density functional theory, both requires the iterative solution of a set of Schr\textbackslash "odinger-like equations. The speed of convergence of such self-consistent field process depends on the complexity of the system under investigation and on the initial guess for the density matrix. An initial density matrix close to the ground-state one will effectively allow one to cut out many of the self-consistent steps necessary to achieve convergence. Here, we predict the density matrix of Kohn-Sham density functional theory by constructing a neural network, which uses the atomic positions as only information. Such neural network provides an initial guess for the density matrix far superior to any other recipes available, in particular for molecules with metallic bonds. Furthermore, the quality of such neural-network density matrix is good enough for the evaluation of interatomic forces. This allows us to run accelerated \{\textbackslash it ab-initio\} molecular dynamics with little to no self-consistent steps.},
- pubstate = {preprint},
- keywords = {\_tablet,alternative approaches,alternative for ML-DFT,AML,BLYP,charge density,density,density matrix,DFT,DFT speedup,DFT speedup with ML,DIIS,HFT,hybrid AI/simulation,initial guess,invariance,MD,ML,ML-Density,ML-DFT,ML-ESM,ML-WFT,molecules,not spin-dependent,prediction of density matrix,prediction of electron density,PySCF,RDMFT,SCF,surrogate model},
+ pubstate = {prepublished},
+ keywords = {\_tablet,alternative approaches,alternative for ML-DFT,AML,BLYP,charge density,density,density matrix,DFT,DFT speedup,DFT speedup with ML,DIIS,HFT,hybrid AI/simulation,initial guess,invariance,Jacobi-Legendre,MD,ML,ML-Density,ML-DFT,ML-ESM,ML-WFT,molecules,not spin-dependent,prediction of density matrix,prediction of electron density,PySCF,RDMFT,SCF,surrogate model},
file = {/Users/wasmer/Nextcloud/Zotero/Hazra et al_2024_Predicting The One-Particle Density Matrix With Machine Learning.pdf;/Users/wasmer/Zotero/storage/LVFIN4XI/2401.html}
}
+@article{hazraPredictingOneParticleDensity2024a,
+ title = {Predicting the {{One-Particle Density Matrix}} with {{Machine Learning}}},
+ author = {Hazra, S. and Patil, U. and Sanvito, S.},
+ date = {2024-06-11},
+ journaltitle = {Journal of Chemical Theory and Computation},
+ shortjournal = {J. Chem. Theory Comput.},
+ volume = {20},
+ number = {11},
+ pages = {4569--4578},
+ publisher = {American Chemical Society},
+ issn = {1549-9618},
+ doi = {10.1021/acs.jctc.4c00042},
+ url = {https://doi.org/10.1021/acs.jctc.4c00042},
+ urldate = {2024-07-05},
+ abstract = {Two of the most widely used electronic-structure theory methods, namely, Hartree–Fock and Kohn–Sham density functional theory, require the iterative solution of a set of Schrödinger-like equations. The speed of convergence of such a process depends on the complexity of the system under investigation, the self-consistent-field algorithm employed, and the initial guess for the density matrix. An initial density matrix close to the ground-state matrix will effectively allow one to cut out many of the self-consistent steps necessary to achieve convergence. Here, we predict the density matrix of Kohn–Sham density functional theory by constructing a neural network that uses only the atomic positions as information. Such a neural network provides an initial guess for the density matrix far superior to that of any other recipes available. Furthermore, the quality of such a neural-network density matrix is good enough for the evaluation of interatomic forces. This allows us to run accelerated ab initio molecular dynamics with little to no self-consistent steps.},
+ keywords = {/unread,alternative approaches,alternative for ML-DFT,AML,BLYP,charge density,density,density matrix,DFT,DFT speedup,DFT speedup with ML,DIIS,HFT,hybrid AI/simulation,initial guess,invariance,Jacobi-Legendre,MD,ML,ML-Density,ML-DFT,ML-ESM,ML-WFT,molecules,not spin-dependent,prediction of density matrix,prediction of electron density,PySCF,RDMFT,SCF,surrogate model},
+ file = {/Users/wasmer/Nextcloud/Zotero/Hazra et al_2024_Predicting the One-Particle Density Matrix with Machine Learning2.pdf}
+}
+
@article{hegdeMachinelearnedApproximationsDensity2017,
title = {Machine-Learned Approximations to {{Density Functional Theory Hamiltonians}}},
author = {Hegde, Ganesh and Bowen, R. Chris},
@@ -6900,13 +7519,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Hegde, Vinay I. and Borg, Christopher K. H. and family=Rosario, given=Zachary, prefix=del, useprefix=true and Kim, Yoolhee and Hutchinson, Maxwell and Antono, Erin and Ling, Julia and Saxe, Paul and Saal, James E. and Meredig, Bryce},
date = {2022-11-05},
eprint = {2007.01988},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2007.01988},
url = {http://arxiv.org/abs/2007.01988},
urldate = {2023-08-19},
abstract = {A central challenge in high throughput density functional theory (HT-DFT) calculations is selecting a combination of input parameters and post-processing techniques that can be used across all materials classes, while also managing accuracy-cost tradeoffs. To investigate the effects of these parameter choices, we consolidate three large HT-DFT databases: Automatic-FLOW (AFLOW), the Materials Project (MP), and the Open Quantum Materials Database (OQMD), and compare reported properties across each pair of databases for materials calculated using the same initial crystal structure. We find that HT-DFT formation energies and volumes are generally more reproducible than band gaps and total magnetizations; for instance, a notable fraction of records disagree on whether a material is metallic (up to 7\%) or magnetic (up to 15\%). The variance between calculated properties is as high as 0.105 eV/atom (median relative absolute difference, or MRAD, of 6\%) for formation energy, 0.65 \{\textbackslash AA\}\$\textasciicircum 3\$/atom (MRAD of 4\%) for volume, 0.21 eV (MRAD of 9\%) for band gap, and 0.15 \$\textbackslash mu\_\{\textbackslash rm B\}\$/formula unit (MRAD of 8\%) for total magnetization, comparable to the differences between DFT and experiment. We trace some of the larger discrepancies to choices involving pseudopotentials, the DFT+U formalism, and elemental reference states, and argue that further standardization of HT-DFT would be beneficial to reproducibility.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AFLOWLIB,Citrine Informatics,DFT,magnetization,materials database,materials project,OQMD,reproducibility,todo-tagging,uncertainty quantification},
file = {/Users/wasmer/Nextcloud/Zotero/Hegde et al_2022_Quantifying uncertainty in high-throughput density functional theory.pdf;/Users/wasmer/Zotero/storage/CCU4ZPKZ/2007.html}
}
@@ -6945,7 +7564,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
urldate = {2021-10-19},
abstract = {Data analyses based on linear methods constitute the simplest, most robust, and transparent approaches to the automatic processing of large amounts of data for building supervised or unsupervised machine learning models. Principal covariates regression (PCovR) is an underappreciated method that interpolates between principal component analysis and linear regression and can be used conveniently to reveal structure-property relations in terms of simple-to-interpret, low-dimensional maps. Here we provide a pedagogic overview of these data analysis schemes, including the use of the kernel trick to introduce an element of non-linearity while maintaining most of the convenience and the simplicity of linear approaches. We then introduce a kernelized version of PCovR and a sparsified extension, and demonstrate the performance of this approach in revealing and predicting structure-property relations in chemistry and materials science, showing a variety of examples including elemental carbon, porous silicate frameworks, organic molecules, amino acid conformers, and molecular materials.},
langid = {english},
- keywords = {\_tablet,KPCovR,KRR,models,original publication,PCovR,regression,Supervised learning,unsupervised learning},
+ keywords = {KPCovR,KRR,models,original publication,PCovR,regression,Supervised learning,unsupervised learning},
file = {/Users/wasmer/Nextcloud/Zotero/Helfrecht et al_2020_Structure-property maps with Kernel principal covariates regression.pdf}
}
@@ -6963,7 +7582,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
url = {https://link.aps.org/doi/10.1103/PhysRevLett.109.076801},
urldate = {2022-05-13},
abstract = {First-principles and model calculations show that the Dirac surface state of the topological insulator Bi2Te3 survives upon moderate Mn doping of the surface layers but can lose its topological character as a function of magnetization direction. The dispersion depends considerably on the direction of the Mn magnetization: for perpendicular magnetization, a gap of 16 meV opens up at the Dirac point; for in-plane magnetization, a tiny gap can be opened or closed in dependence on the magnetization azimuth. The ground state is ferromagnetic, with a critical temperature of 12 K. The results provide a path towards a magnetic control of the topological character of the Dirac surface state and its consequences to spin-dependent transport properties.},
- keywords = {\_tablet,AFM,Bi2Te3,bismuth telluride,Heisenberg model,Jij,KKR,magnetic doping,Mn doping,rec-by-ruess,topological insulator},
+ keywords = {AFM,Bi2Te3,bismuth telluride,Heisenberg model,Jij,KKR,magnetic doping,Mn doping,rec-by-ruess,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Henk et al_2012_Topological Character and Magnetism of the Dirac State in Mn-Doped Bi2Te3.pdf;/Users/wasmer/Zotero/storage/W6BV33VI/Henk et al_2012_Topological Character and Magnetism of the Dirac State in Mn-Doped Bi2Te3.pdf;/Users/wasmer/Zotero/storage/ZTFJBVIM/PhysRevLett.109.html}
}
@@ -7014,7 +7633,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
volume = {33},
number = {8},
eprint = {2009.01665},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
pages = {085503},
issn = {0953-8984, 1361-648X},
@@ -7056,7 +7675,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
shortjournal = {Faraday Discuss.},
volume = {224},
eprint = {2004.13549},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
pages = {227--246},
issn = {1359-6640, 1364-5498},
@@ -7077,7 +7696,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
volume = {153},
number = {5},
eprint = {2005.05848},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
pages = {054114},
issn = {0021-9606, 1089-7690},
@@ -7097,7 +7716,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
shortjournal = {Journal of Computational Physics},
volume = {459},
eprint = {2109.14018},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
pages = {111127},
issn = {00219991},
@@ -7114,7 +7733,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Herbst, Michael F. and Wessel, Stefan and Rizzi, Matteo and Stamm, Benjamin},
date = {2021-11-24},
eprint = {2110.15665},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:quant-ph},
url = {http://arxiv.org/abs/2110.15665},
urldate = {2022-01-11},
@@ -7197,7 +7816,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
urldate = {2023-06-14},
abstract = {Topological insulators are emergent states of quantum matter that are gapped in the bulk with time-reversal symmetry-preserved gapless edge/surface states, adiabatically distinct from conventional materials. By proximity to various magnets and superconductors, topological insulators show novel physics at the interfaces, which give rise to two new areas named topological spintronics and topological quantum computation. Effects in the former such as the spin torques, spin-charge conversion, topological antiferromagnetic spintronics, and skyrmions realized in topological systems will be addressed. In the latter, a superconducting pairing gap leads to a state that supports Majorana fermions states, which may provide a new path for realizing topological quantum computation. Various signatures of Majorana zero modes/edge mode in topological superconductors will be discussed. The review ends by outlooks and potential applications of topological insulators. Topological superconductors that are fabricated using topological insulators with superconductors have a full pairing gap in the bulk and gapless surface states consisting of Majorana fermions. The theory of topological superconductors is reviewed, in close analogy to the theory of topological insulators.},
langid = {english},
- keywords = {\_tablet,applications,ARPES,Hall effect,Hall QAHE,Majorana,quantum computing,review,skyrmions,Spintronics,topological insulator,Topological Superconductor},
+ keywords = {applications,ARPES,Hall effect,Hall QAHE,Majorana,quantum computing,review,skyrmions,Spintronics,topological insulator,Topological Superconductor},
file = {/Users/wasmer/Nextcloud/Zotero/He et al_2019_Topological insulator.pdf}
}
@@ -7210,7 +7829,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
volume = {108},
number = {7},
eprint = {2303.11207},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, physics:quant-ph},
pages = {075152},
issn = {2469-9950, 2469-9969},
@@ -7260,6 +7879,38 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
file = {/Users/wasmer/Nextcloud/Zotero/Hicks et al_2018_AFLOW-SYM.pdf}
}
+@online{hilgersApplicationBatchLearning2023,
+ title = {Application of Batch Learning for Boosting High-Throughput Ab Initio Success Rates and Reducing Computational Effort Required Using Data-Driven Processes},
+ author = {Hilgers, Robin and Wortmann, Daniel and Blügel, Stefan},
+ date = {2023-11-26},
+ eprint = {2311.15430},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat},
+ doi = {10.48550/arXiv.2311.15430},
+ url = {http://arxiv.org/abs/2311.15430},
+ urldate = {2024-06-05},
+ abstract = {The increased availability of computing time, in recent years, allows for systematic high-throughput studies of material classes with the purpose of both screening for materials with remarkable properties and understanding how structural configuration and material composition affect macroscopic attributes manifestation. However, when conducting systematic high-throughput studies, the individual ab initio calculations' success depends on the quality of the chosen input quantities. On a large scale, improving input parameters by trial and error is neither efficient nor systematic. We present a systematic, high-throughput compatible, and machine learning-based approach to improve the input parameters optimized during a DFT computation or workflow. This approach of integrating machine learning into a typical high-throughput workflow demonstrates the advantages and necessary considerations for a systematic study of magnetic multilayers of 3\$d\$ transition metal layers on FCC noble metal substrates. For 6660 film systems, we were able to improve the overall success rate of our high-throughput FLAPW-based structural relaxations from \$64.8 \textbackslash\%\$ to \$94.3\textbackslash{} \textbackslash\%\$ while at the same time requiring \$17\textbackslash{} \textbackslash\%\$ less computational time for each successful relaxation.},
+ pubstate = {prepublished},
+ keywords = {active learning,AiiDA,AiiDA-FLEUR,AML,DFT,DVC,FLEUR,FZJ,HTC,ML,PGI,PGI-1/IAS-1,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Hilgers et al_2023_Application of batch learning for boosting high-throughput ab initio success.pdf;/Users/wasmer/Zotero/storage/IBUSAKTA/2311.html}
+}
+
+@online{hilgersMachineLearningbasedEstimation2023,
+ title = {Machine {{Learning-based}} Estimation and Explainable Artificial Intelligence-Supported Interpretation of the Critical Temperature from Magnetic Ab Initio {{Heusler}} Alloys Data},
+ author = {Hilgers, Robin and Wortmann, Daniel and Blügel, Stefan},
+ date = {2023-11-26},
+ eprint = {2311.15423},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat},
+ doi = {10.48550/arXiv.2311.15423},
+ url = {http://arxiv.org/abs/2311.15423},
+ urldate = {2024-06-05},
+ abstract = {Machine Learning (ML) has impacted numerous areas of materials science, most prominently improving molecular simulations, where force fields were trained on previously relaxed structures. One natural next step is to predict material properties beyond structure. In this work, we investigate the applicability and explainability of ML methods in the use case of estimating the critical temperature for magnetic Heusler alloys calculated using ab initio methods determined materials-specific magnetic interactions and a subsequent Monte Carlo (MC) approach. We compare the performance of regression and classification models to predict the range of the critical temperature of given compounds without performing the MC calculations. Since the MC calculation requires computational resources in the same order of magnitude as the density-functional theory (DFT) calculation, it would be advantageous to replace either step with a less computationally intensive method such as ML. We discuss the necessity to generate the magnetic ab initio results to make a quantitative prediction of the critical temperature. We used state-of-the-art explainable artificial intelligence (XAI) methods to extract physical relations and deepen our understanding of patterns learned by our models from the examined data.},
+ pubstate = {prepublished},
+ keywords = {AML,Curie temperature,DFT,FLEUR,FZJ,Heusler,magnetism,ML,PGI,PGI-1/IAS-1,SHAP},
+ file = {/Users/wasmer/Nextcloud/Zotero/Hilgers et al_2023_Machine Learning-based estimation and explainable artificial.pdf;/Users/wasmer/Zotero/storage/9ILANUED/2311.html}
+}
+
@article{himanenDScribeLibraryDescriptors2020,
title = {{{DScribe}}: {{Library}} of Descriptors for Machine Learning in Materials Science},
shorttitle = {{{DScribe}}},
@@ -7337,13 +7988,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Hodapp, Max and Shapeev, Alexander},
date = {2023-04-17},
eprint = {2304.08226},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2304.08226},
url = {http://arxiv.org/abs/2304.08226},
urldate = {2023-05-26},
abstract = {Machine-learning interatomic potentials (MLIPs) have made a significant contribution to the recent progress in the fields of computational materials and chemistry due to MLIPs' ability of accurately approximating energy landscapes of quantum-mechanical models while being orders of magnitude more computationally efficient. However, the computational cost and number of parameters of many state-of-the-art MLIPs increases exponentially with the number of atomic features. Tensor (non-neural) networks, based on low-rank representations of high-dimensional tensors, have been a way to reduce the number of parameters in approximating multidimensional functions, however, it is often not easy to encode the model symmetries into them. In this work we develop a formalism for rank-efficient equivariant tensor networks (ETNs), i.e., tensor networks that remain invariant under actions of SO(3) upon contraction. All the key algorithms of tensor networks like orthogonalization of cores and DMRG carry over to our equivariant case. Moreover, we show that many elements of modern neural network architectures like message passing, pulling, or attention mechanisms, can in some form be implemented into the ETNs. Based on ETNs, we develop a new class of polynomial-based MLIPs that demonstrate superior performance over existing MLIPs for multicomponent systems.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ACE,AML,binary systems,dimensionality reduction,DMRG,equivariant,HEA,ML,MLP,MTP,multi-component systems,QM7,QM9,quaternary systems,SO(3),tensor network},
file = {/Users/wasmer/Nextcloud/Zotero/Hodapp_Shapeev_2023_Equivariant Tensor Networks.pdf;/Users/wasmer/Zotero/storage/J2HAIBJ3/2304.html}
}
@@ -7425,6 +8076,22 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
file = {/Users/wasmer/Zotero/storage/Q7HV8D2L/intro.html}
}
+@online{holzmullerFrameworkBenchmarkDeep2023,
+ title = {A {{Framework}} and {{Benchmark}} for {{Deep Batch Active Learning}} for {{Regression}}},
+ author = {Holzmüller, David and Zaverkin, Viktor and Kästner, Johannes and Steinwart, Ingo},
+ date = {2023-08-01},
+ eprint = {2203.09410},
+ eprinttype = {arXiv},
+ eprintclass = {cs, stat},
+ doi = {10.48550/arXiv.2203.09410},
+ url = {http://arxiv.org/abs/2203.09410},
+ urldate = {2024-06-15},
+ abstract = {The acquisition of labels for supervised learning can be expensive. To improve the sample efficiency of neural network regression, we study active learning methods that adaptively select batches of unlabeled data for labeling. We present a framework for constructing such methods out of (network-dependent) base kernels, kernel transformations, and selection methods. Our framework encompasses many existing Bayesian methods based on Gaussian process approximations of neural networks as well as non-Bayesian methods. Additionally, we propose to replace the commonly used last-layer features with sketched finite-width neural tangent kernels and to combine them with a novel clustering method. To evaluate different methods, we introduce an open-source benchmark consisting of 15 large tabular regression data sets. Our proposed method outperforms the state-of-the-art on our benchmark, scales to large data sets, and works out-of-the-box without adjusting the network architecture or training code. We provide open-source code that includes efficient implementations of all kernels, kernel transformations, and selection methods, and can be used for reproducing our results.},
+ pubstate = {prepublished},
+ keywords = {/unread,active learning,Bayesian methods,Deep learning,General ML,GPR,kernel methods,library,ML,neural network,regression,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Holzmüller et al_2023_A Framework and Benchmark for Deep Batch Active Learning for Regression.pdf;/Users/wasmer/Zotero/storage/FKKA4ZKI/2203.html}
+}
+
@article{hongReducingTimeDiscovery2021,
title = {Reducing {{Time}} to {{Discovery}}: {{Materials}} and {{Molecular Modeling}}, {{Imaging}}, {{Informatics}}, and {{Integration}}},
shorttitle = {Reducing {{Time}} to {{Discovery}}},
@@ -7459,7 +8126,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
url = {https://doi.org/10.1063/5.0012407},
urldate = {2023-05-06},
abstract = {We present a machine learning approach for accurately predicting formation energies of binary compounds in the context of crystal structure predictions. The success of any machine learning model depends significantly on the choice of representation used to encode the relevant physical information into machine-learnable data. We test different representation schemes based on partial radial and angular distribution functions (RDF+ADF) on Al–Ni and Cd–Te structures generated using our genetic algorithm for structure prediction. We observe a remarkable improvement in predictive accuracy upon transitioning from global to atom-centered representations, resulting in a threefold decrease in prediction errors. We show that a support vector regression model using a combination of atomic radial and angular distribution functions performs best at the formation energy prediction task, providing small root mean squared errors of 3.9\,meV/atom and 10.9\,meV/atom for Al–Ni and Cd–Te, respectively. We test the performance of our models against common traditional descriptors and find that RDF- and ADF-based representations significantly outperform many of those in the prediction of formation energies. The high accuracy of predictions makes our machine learning models great candidates for the exploration of energy landscapes.},
- keywords = {/unread,\_tablet,ACDC,ACSF,ADF descriptor,AML,benchmarking,binary systems,CFID,Coulomb matrix,crystal structure prediction,descriptor comparison,descriptors,LAMMPS,MBTR,MD,ML,OFM descriptor,PES,prediction of energy,prediction of formation energy,RDF descriptor,SB descriptors,SOAP,structure prediction},
+ keywords = {/unread,ACDC,ACSF,ADF descriptor,AML,benchmarking,binary systems,CFID,Coulomb matrix,crystal structure prediction,descriptor comparison,descriptors,LAMMPS,MBTR,MD,ML,OFM descriptor,PES,prediction of energy,prediction of formation energy,RDF descriptor,SB descriptors,SOAP,structure prediction},
file = {/Users/wasmer/Nextcloud/Zotero/Honrao et al_2020_Augmenting machine learning of energy landscapes with local structural.pdf;/Users/wasmer/Zotero/storage/VFKSDW8H/Augmenting-machine-learning-of-energy-landscapes.html}
}
@@ -7468,13 +8135,13 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
author = {Hoogeboom, Emiel and Satorras, Victor Garcia and Vignac, Clément and Welling, Max},
date = {2022-06-16},
eprint = {2203.17003},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, q-bio, stat},
doi = {10.48550/arXiv.2203.17003},
url = {http://arxiv.org/abs/2203.17003},
urldate = {2023-08-22},
abstract = {This work introduces a diffusion model for molecule generation in 3D that is equivariant to Euclidean transformations. Our E(3) Equivariant Diffusion Model (EDM) learns to denoise a diffusion process with an equivariant network that jointly operates on both continuous (atom coordinates) and categorical features (atom types). In addition, we provide a probabilistic analysis which admits likelihood computation of molecules using our model. Experimentally, the proposed method significantly outperforms previous 3D molecular generative methods regarding the quality of generated samples and efficiency at training time.},
- pubstate = {preprint},
+ pubstate = {prepublished},
version = {2},
keywords = {diffusion model,generative models,todo-tagging,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Hoogeboom et al_2022_Equivariant Diffusion for Molecule Generation in 3D.pdf;/Users/wasmer/Zotero/storage/K6KWYTSV/2203.html}
@@ -7495,6 +8162,42 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
file = {/Users/wasmer/Nextcloud/Zotero/Horsch et al_2023_European standardization efforts from FAIR toward explainable-AI-ready data.pdf;/Users/wasmer/Zotero/storage/ETXXB8B6/10193944.html}
}
+@article{hortonPromisesPerilsComputational2021,
+ title = {Promises and Perils of Computational Materials Databases},
+ author = {Horton, M. K. and Dwaraknath, S. and Persson, K. A.},
+ date = {2021-01},
+ journaltitle = {Nature Computational Science},
+ shortjournal = {Nat Comput Sci},
+ volume = {1},
+ number = {1},
+ pages = {3--5},
+ publisher = {Nature Publishing Group},
+ issn = {2662-8457},
+ doi = {10.1038/s43588-020-00016-5},
+ url = {https://www.nature.com/articles/s43588-020-00016-5},
+ urldate = {2024-05-21},
+ abstract = {Over the past decade, the materials science community has fostered the development of materials databases from high-performance computation. While these databases have achieved great success, there are still several challenges to be addressed for the community to realize the full potential of the materials-by-design era.},
+ langid = {english},
+ keywords = {/unread,AFLOW,commentary,computational database,computational materials database,Database,materials,Materials Cloud,materials database,materials project,NOMAD,OQMD,perspective},
+ file = {/Users/wasmer/Nextcloud/Zotero/Horton et al_2021_Promises and perils of computational materials databases.pdf}
+}
+
+@online{houPhysicsinformedActiveLearning2024,
+ title = {Physics-Informed Active Learning for Accelerating Quantum Chemical Simulations},
+ author = {Hou, Yi-Fan and Zhang, Lina and Zhang, Quanhao and Ge, Fuchun and Dral, Pavlo O.},
+ date = {2024-04-17},
+ eprint = {2404.11811},
+ eprinttype = {arXiv},
+ eprintclass = {physics},
+ doi = {10.48550/arXiv.2404.11811},
+ url = {http://arxiv.org/abs/2404.11811},
+ urldate = {2024-07-03},
+ abstract = {Quantum chemical simulations can be greatly accelerated by constructing machine learning potentials, which is often done using active learning (AL). The usefulness of the constructed potentials is often limited by the high effort required and their insufficient robustness in the simulations. Here we introduce the end-to-end AL for constructing robust data-efficient potentials with affordable investment of time and resources and minimum human interference. Our AL protocol is based on the physics-informed sampling of training points, automatic selection of initial data, and uncertainty quantification. The versatility of this protocol is shown in our implementation of quasi-classical molecular dynamics for simulating vibrational spectra, conformer search of a key biochemical molecule, and time-resolved mechanism of the Diels-Alder reaction. These investigations took us days instead of weeks of pure quantum chemical calculations on a high-performance computing cluster.},
+ pubstate = {prepublished},
+ keywords = {/unread,active learning,AML,GAP,KRR,library,MACE,MEGNet,ML,MLatom,MLP,NequIP,PhysNet,SchNet,SOAP,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Hou et al_2024_Physics-informed active learning for accelerating quantum chemical simulations.pdf;/Users/wasmer/Zotero/storage/BSLJJWPZ/2404.html}
+}
+
@article{huAisNetUniversalInteratomic2023,
title = {{{AisNet}}: {{A Universal Interatomic Potential Neural Network}} with {{Encoded Local Environment Features}}},
shorttitle = {{{AisNet}}},
@@ -7511,7 +8214,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
url = {https://doi.org/10.1021/acs.jcim.3c00077},
urldate = {2023-06-30},
abstract = {This paper proposes a new interatomic potential energy neural network, AisNet, which can efficiently predict atomic energies and forces covering different molecular and crystalline materials by encoding universal local environment features, such as elements and atomic positions. Inspired by the framework of SchNet, AisNet consists of an encoding module combining autoencoder with embedding, the triplet loss function and an atomic central symmetry function (ACSF), an interaction module with a periodic boundary condition (PBC), and a prediction module. In molecules, the prediction accuracy of AisNet is comparabel with SchNet on the MD17 dataset, mainly attributed to the effective capture of chemical functional groups through the interaction module. In selected metal and ceramic material datasets, the introduction of ACSF improves the overall accuracy of AisNet by an average of 16.8\% for energy and 28.6\% for force. Furthermore, a close relationship is found between the feature ratio (i.e., ACSF and embedding) and the force prediction errors, exhibiting similar spoon-shaped curves in the datasets of Cu and HfO2. AisNet produces highly accurate predictions in single-commponent alloys with little data, suggesting the encoding process reduces dependence on the number and richness of datasets. Especially for force prediction, AisNet exceeds SchNet by 19.8\% for Al and even 81.2\% higher than DeepMD on a ternary FeCrAl alloy. Capable of processing multivariate features, our model is likely to be applied to a wider range of material systems by incorporating more atomic descriptions.},
- keywords = {/unread,AML,ML,MLP,universal potential}
+ keywords = {/unread,ACSF,AML,autoencoder,DeePMD-kit,descriptors,embedding,GNN,ML,MLP,SchNet,with-code}
}
@article{huangCentralRoleDensity2023,
@@ -7549,6 +8252,23 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
file = {/Users/wasmer/Nextcloud/Zotero/Huang et al_2017_Emerging topological states in quasi-two-dimensional materials.pdf;/Users/wasmer/Zotero/storage/R26FLUTA/wcms.html}
}
+@article{huangProvablyEfficientMachine2022,
+ title = {Provably Efficient Machine Learning for Quantum Many-Body Problems},
+ author = {Huang, Hsin-Yuan and Kueng, Richard and Torlai, Giacomo and Albert, Victor V. and Preskill, John},
+ date = {2022-09-23},
+ journaltitle = {Science},
+ volume = {377},
+ number = {6613},
+ pages = {eabk3333},
+ publisher = {American Association for the Advancement of Science},
+ doi = {10.1126/science.abk3333},
+ url = {https://www.science.org/doi/10.1126/science.abk3333},
+ urldate = {2024-06-06},
+ abstract = {Classical machine learning (ML) provides a potentially powerful approach to solving challenging quantum many-body problems in physics and chemistry. However, the advantages of ML over traditional methods have not been firmly established. In this work, we prove that classical ML algorithms can efficiently predict ground-state properties of gapped Hamiltonians after learning from other Hamiltonians in the same quantum phase of matter. By contrast, under a widely accepted conjecture, classical algorithms that do not learn from data cannot achieve the same guarantee. We also prove that classical ML algorithms can efficiently classify a wide range of quantum phases. Extensive numerical experiments corroborate our theoretical results in a variety of scenarios, including Rydberg atom systems, two-dimensional random Heisenberg models, symmetry-protected topological phases, and topologically ordered phases.},
+ keywords = {/unread,classical ML,classical shadow,ML,ML-FPO,ML-QMBP,prediction of ground-state properties,quantum computing,Quantum simulation,quantum state tomography},
+ file = {/Users/wasmer/Nextcloud/Zotero/Huang et al_2022_Provably efficient machine learning for quantum many-body problems.pdf}
+}
+
@article{huangUnveilingComplexStructureproperty2023,
title = {Unveiling the Complex Structure-Property Correlation of Defects in {{2D}} Materials Based on High Throughput Datasets},
author = {Huang, Pengru and Lukin, Ruslan and Faleev, Maxim and Kazeev, Nikita and Al-Maeeni, Abdalaziz Rashid and Andreeva, Daria V. and Ustyuzhanin, Andrey and Tormasov, Alexander and Castro Neto, A. H. and Novoselov, Kostya S.},
@@ -7566,7 +8286,7 @@ Subject\_term\_id: condensed-matter-physics;theory-and-computation},
abstract = {Modification of physical properties of materials and design of materials with on-demand characteristics is at the heart of modern technology. Rare application relies on pure materials—most devices and technologies require careful design of materials properties through alloying, creating heterostructures of composites, or controllable introduction of defects. At the same time, such designer materials are notoriously difficult to model. Thus, it is very tempting to apply machine learning methods to such systems. Unfortunately, there is only a handful of machine learning-friendly material databases available these days. We develop a platform for easy implementation of machine learning techniques to materials design and populate it with datasets on pristine and defected materials. Here we introduce the 2D Material Defect (2DMD) datasets that include defect properties of represented 2D materials such as MoS2, WSe2, hBN, GaSe, InSe, and black phosphorous, calculated using DFT. Our study provides a data-driven physical understanding of complex behaviors of defect properties in 2D materials, holding promise for a guide to the development of efficient machine learning models. In addition, with the increasing enrollment of datasets, our database could provide a platform for designing materials with predetermined properties.},
issue = {1},
langid = {english},
- keywords = {\_tablet,2D material,2DMD dataset,AML,Database,database generation,defect descriptor,defect screening,defects,disordered,high-density defects,High-throughput,HTC,materials screening,ML,MoS2,PBE,physics,point defects,sampling,spin-polarized,TMDC,vacancies,VASP,with-data},
+ keywords = {2D material,2DMD dataset,AML,Database,database generation,defect descriptor,defect screening,defects,disordered,high-density defects,High-throughput,HTC,materials screening,ML,MoS2,PBE,physics,point defects,sampling,spin-polarized,TMDC,vacancies,VASP,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Huang et al_2023_Unveiling the complex structure-property correlation of defects in 2D materials.pdf}
}
@@ -7602,7 +8322,7 @@ Subject\_term\_id: computational-methods;research-management},
author = {Huber, Sebastiaan P. and Bosoni, Emanuele and Bercx, Marnik and Bröder, Jens and Degomme, Augustin and Dikan, Vladimir and Eimre, Kristjan and Flage-Larsen, Espen and Garcia, Alberto and Genovese, Luigi and Gresch, Dominik and Johnston, Conrad and Petretto, Guido and Poncé, Samuel and Rignanese, Gian-Marco and Sewell, Christopher J. and Smit, Berend and Tseplyaev, Vasily and Uhrin, Martin and Wortmann, Daniel and Yakutovich, Aliaksandr V. and Zadoks, Austin and Zarabadi-Poor, Pezhman and Zhu, Bonan and Marzari, Nicola and Pizzi, Giovanni},
date = {2021-05-11},
eprint = {2105.05063},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
url = {http://arxiv.org/abs/2105.05063},
urldate = {2021-06-23},
@@ -7637,30 +8357,46 @@ Subject\_term\_id: computational-methods;research-management},
author = {Huguenin-Dumittan, Kevin K. and Loche, Philip and Haoran, Ni and Ceriotti, Michele},
date = {2023-08-25},
eprint = {2308.13208},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2308.13208},
url = {http://arxiv.org/abs/2308.13208},
urldate = {2023-08-29},
abstract = {Most of the existing machine-learning schemes applied to atomic-scale simulations rely on a local description of the geometry of a structure, and struggle to model effects that are driven by long-range physical interactions. Efforts to overcome these limitations have focused on the direct incorporation of electrostatics, which is the most prominent effect, often relying on architectures that mirror the functional form of explicit physical models. Including other forms of non-bonded interactions, or predicting properties other than the interatomic potential, requires ad hoc modifications. We propose an alternative approach that extends the long-distance equivariant (LODE) framework to generate local descriptors of an atomic environment that resemble non-bonded potentials with arbitrary asymptotic behaviors, ranging from point-charge electrostatics to dispersion forces. We show that the LODE formalism is amenable to a direct physical interpretation in terms of a generalized multipole expansion, that simplifies its implementation and reduces the number of descriptors needed to capture a given asymptotic behavior. These generalized LODE features provide improved extrapolation capabilities when trained on structures dominated by a given asymptotic behavior, but do not help in capturing the wildly different energy scales that are relevant for a more heterogeneous data set. This approach provides a practical scheme to incorporate different types of non-bonded interactions, and a framework to investigate the interplay of physical and data-related considerations that underlie this challenging modeling problem.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AML,descriptors,electrostatic interaction,library,LODE,long-range interaction,ML,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Huguenin-Dumittan et al_2023_Physics-inspired Equivariant Descriptors of Non-bonded Interactions.pdf;/Users/wasmer/Zotero/storage/H7WQMUXS/2308.html}
}
+@online{huhPlatonicRepresentationHypothesis2024,
+ title = {The {{Platonic Representation Hypothesis}}},
+ author = {Huh, Minyoung and Cheung, Brian and Wang, Tongzhou and Isola, Phillip},
+ date = {2024-05-13},
+ eprint = {2405.07987},
+ eprinttype = {arXiv},
+ eprintclass = {cs},
+ doi = {10.48550/arXiv.2405.07987},
+ url = {http://arxiv.org/abs/2405.07987},
+ urldate = {2024-05-29},
+ abstract = {We argue that representations in AI models, particularly deep networks, are converging. First, we survey many examples of convergence in the literature: over time and across multiple domains, the ways by which different neural networks represent data are becoming more aligned. Next, we demonstrate convergence across data modalities: as vision models and language models get larger, they measure distance between datapoints in a more and more alike way. We hypothesize that this convergence is driving toward a shared statistical model of reality, akin to Plato's concept of an ideal reality. We term such a representation the platonic representation and discuss several possible selective pressures toward it. Finally, we discuss the implications of these trends, their limitations, and counterexamples to our analysis.},
+ pubstate = {prepublished},
+ keywords = {/unread,Computer Science - Artificial Intelligence,Computer Science - Computer Vision and Pattern Recognition,Computer Science - Machine Learning,Computer Science - Neural and Evolutionary Computing},
+ file = {/Users/wasmer/Nextcloud/Zotero/Huh et al_2024_The Platonic Representation Hypothesis.pdf;/Users/wasmer/Zotero/storage/R7C9FVMV/2405.html}
+}
+
@online{huOGBLSCLargeScaleChallenge2021,
title = {{{OGB-LSC}}: {{A Large-Scale Challenge}} for {{Machine Learning}} on {{Graphs}}},
shorttitle = {{{OGB-LSC}}},
author = {Hu, Weihua and Fey, Matthias and Ren, Hongyu and Nakata, Maho and Dong, Yuxiao and Leskovec, Jure},
date = {2021-10-20},
eprint = {2103.09430},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2103.09430},
url = {http://arxiv.org/abs/2103.09430},
urldate = {2023-07-24},
abstract = {Enabling effective and efficient machine learning (ML) over large-scale graph data (e.g., graphs with billions of edges) can have a great impact on both industrial and scientific applications. However, existing efforts to advance large-scale graph ML have been largely limited by the lack of a suitable public benchmark. Here we present OGB Large-Scale Challenge (OGB-LSC), a collection of three real-world datasets for facilitating the advancements in large-scale graph ML. The OGB-LSC datasets are orders of magnitude larger than existing ones, covering three core graph learning tasks -- link prediction, graph regression, and node classification. Furthermore, we provide dedicated baseline experiments, scaling up expressive graph ML models to the massive datasets. We show that expressive models significantly outperform simple scalable baselines, indicating an opportunity for dedicated efforts to further improve graph ML at scale. Moreover, OGB-LSC datasets were deployed at ACM KDD Cup 2021 and attracted more than 500 team registrations globally, during which significant performance improvements were made by a variety of innovative techniques. We summarize the common techniques used by the winning solutions and highlight the current best practices in large-scale graph ML. Finally, we describe how we have updated the datasets after the KDD Cup to further facilitate research advances. The OGB-LSC datasets, baseline code, and all the information about the KDD Cup are available at https://ogb.stanford.edu/docs/lsc/ .},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,benchmark dataset,benchmarking,Database,graph database,graph ML,OGB,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Hu et al_2021_OGB-LSC.pdf;/Users/wasmer/Zotero/storage/CCNGLUCV/2103.html}
}
@@ -7671,13 +8407,13 @@ Subject\_term\_id: computational-methods;research-management},
author = {Hu, Weihua and Fey, Matthias and Zitnik, Marinka and Dong, Yuxiao and Ren, Hongyu and Liu, Bowen and Catasta, Michele and Leskovec, Jure},
date = {2021-02-24},
eprint = {2005.00687},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.2005.00687},
url = {http://arxiv.org/abs/2005.00687},
urldate = {2023-07-24},
abstract = {We present the Open Graph Benchmark (OGB), a diverse set of challenging and realistic benchmark datasets to facilitate scalable, robust, and reproducible graph machine learning (ML) research. OGB datasets are large-scale, encompass multiple important graph ML tasks, and cover a diverse range of domains, ranging from social and information networks to biological networks, molecular graphs, source code ASTs, and knowledge graphs. For each dataset, we provide a unified evaluation protocol using meaningful application-specific data splits and evaluation metrics. In addition to building the datasets, we also perform extensive benchmark experiments for each dataset. Our experiments suggest that OGB datasets present significant challenges of scalability to large-scale graphs and out-of-distribution generalization under realistic data splits, indicating fruitful opportunities for future research. Finally, OGB provides an automated end-to-end graph ML pipeline that simplifies and standardizes the process of graph data loading, experimental setup, and model evaluation. OGB will be regularly updated and welcomes inputs from the community. OGB datasets as well as data loaders, evaluation scripts, baseline code, and leaderboards are publicly available at https://ogb.stanford.edu .},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,benchmark dataset,benchmarking,Database,graph database,graph ML,library,original publication,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Hu et al_2021_Open Graph Benchmark.pdf;/Users/wasmer/Zotero/storage/L795VLZX/2005.html}
}
@@ -7687,7 +8423,7 @@ Subject\_term\_id: computational-methods;research-management},
author = {Huo, Haoyan and Rupp, Matthias},
date = {2018-01-02},
eprint = {1704.06439},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/1704.06439},
urldate = {2021-06-29},
@@ -7701,13 +8437,13 @@ Subject\_term\_id: computational-methods;research-management},
author = {Hu, Weihua and Liu, Bowen and Gomes, Joseph and Zitnik, Marinka and Liang, Percy and Pande, Vijay and Leskovec, Jure},
date = {2020-02-18},
eprint = {1905.12265},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.1905.12265},
url = {http://arxiv.org/abs/1905.12265},
urldate = {2023-09-25},
abstract = {Many applications of machine learning require a model to make accurate pre-dictions on test examples that are distributionally different from training ones, while task-specific labels are scarce during training. An effective approach to this challenge is to pre-train a model on related tasks where data is abundant, and then fine-tune it on a downstream task of interest. While pre-training has been effective in many language and vision domains, it remains an open question how to effectively use pre-training on graph datasets. In this paper, we develop a new strategy and self-supervised methods for pre-training Graph Neural Networks (GNNs). The key to the success of our strategy is to pre-train an expressive GNN at the level of individual nodes as well as entire graphs so that the GNN can learn useful local and global representations simultaneously. We systematically study pre-training on multiple graph classification datasets. We find that naive strategies, which pre-train GNNs at the level of either entire graphs or individual nodes, give limited improvement and can even lead to negative transfer on many downstream tasks. In contrast, our strategy avoids negative transfer and improves generalization significantly across downstream tasks, leading up to 9.4\% absolute improvements in ROC-AUC over non-pre-trained models and achieving state-of-the-art performance for molecular property prediction and protein function prediction.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AML,GNN,graph ML,ML,pretrained models,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Hu et al_2020_Strategies for Pre-training Graph Neural Networks.pdf;/Users/wasmer/Zotero/storage/C2HIQ9TR/1905.html}
}
@@ -7717,13 +8453,13 @@ Subject\_term\_id: computational-methods;research-management},
author = {Hutchinson, Maxwell L. and Antono, Erin and Gibbons, Brenna M. and Paradiso, Sean and Ling, Julia and Meredig, Bryce},
date = {2017-11-02},
eprint = {1711.05099},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, stat},
doi = {10.48550/arXiv.1711.05099},
url = {http://arxiv.org/abs/1711.05099},
urldate = {2023-08-19},
abstract = {Despite increasing focus on data publication and discovery in materials science and related fields, the global view of materials data is highly sparse. This sparsity encourages training models on the union of multiple datasets, but simple unions can prove problematic as (ostensibly) equivalent properties may be measured or computed differently depending on the data source. These hidden contextual differences introduce irreducible errors into analyses, fundamentally limiting their accuracy. Transfer learning, where information from one dataset is used to inform a model on another, can be an effective tool for bridging sparse data while preserving the contextual differences in the underlying measurements. Here, we describe and compare three techniques for transfer learning: multi-task, difference, and explicit latent variable architectures. We show that difference architectures are most accurate in the multi-fidelity case of mixed DFT and experimental band gaps, while multi-task most improves classification performance of color with band gaps. For activation energies of steps in NO reduction, the explicit latent variable method is not only the most accurate, but also enjoys cancellation of errors in functions that depend on multiple tasks. These results motivate the publication of high quality materials datasets that encode transferable information, independent of industrial or academic interest in the particular labels, and encourage further development and application of transfer learning methods to materials informatics problems.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,chemical reaction,Citrine Informatics,DFT,difference learning,experimental data,materials,materials project,ML,multi-fidelity,multi-task learning,prediction of bandgap,random forest,small data,transfer learning},
file = {/Users/wasmer/Nextcloud/Zotero/Hutchinson et al_2017_Overcoming data scarcity with transfer learning.pdf;/Users/wasmer/Zotero/storage/6XFVNRE9/1711.html}
}
@@ -7808,7 +8544,7 @@ Subject\_term\_id: computational-methods;research-management},
abstract = {A deep neural network method is developed to learn the density functional theory (DFT) Hamiltonian as a function of atomic structure. This approach provides a solution to the accuracy–efficiency dilemma of DFT and opens opportunities to investigate large-scale materials, such as twisted van der Waals materials.},
issue = {7},
langid = {english},
- keywords = {\_tablet,2D material,AML,Berry phase,briefing,Computational methods,DeepH,DFT,e3nn,Electronic properties and materials,Electronic structure,equivariant,GGA,graphene,heterostructures,library,materials,ML,ML-DFT,ML-ESM,MPNN,OpenMX,PBE,prediction of bandstructure,prediction of Berry phase,prediction of Hamiltonian matrix,SOC,twisted bilayer graphene,with-code,with-data},
+ keywords = {2D material,AML,Berry phase,briefing,Computational methods,DeepH,DFT,e3nn,Electronic properties and materials,Electronic structure,equivariant,GGA,graphene,heterostructures,library,materials,ML,ML-DFT,ML-ESM,MPNN,OpenMX,PBE,prediction of bandstructure,prediction of Berry phase,prediction of Hamiltonian matrix,SOC,twisted bilayer graphene,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/2022_Improving the efficiency of ab initio electronic-structure calculations by deep.pdf}
}
@@ -7822,7 +8558,7 @@ Subject\_term\_id: computational-methods;research-management},
urldate = {2023-08-19},
abstract = {Learn about challenges in Machine Learning for Materials. See how Citrine has overcome these challenges and why off-the-shelf open-source AI will require a lot of tailoring to make it work in this space.},
langid = {american},
- keywords = {/unread,AML,Citrine Informatics,compositional descriptors,materials,ML,sequential learning,small data,surrogate model,transfer learning,uncertainty quantification},
+ keywords = {AML,Citrine Informatics,compositional descriptors,materials,ML,sequential learning,small data,surrogate model,transfer learning,uncertainty quantification},
file = {/Users/wasmer/Nextcloud/Zotero/Informatics_2021_Challenges in Machine Learning for Materials - AI White Paper.pdf;/Users/wasmer/Zotero/storage/M5PWI97V/white-paper-challenges-in-machine-learning-for-materials.html}
}
@@ -7846,6 +8582,20 @@ Subject\_term\_id: computational-methods;research-management},
file = {/Users/wasmer/Nextcloud/Zotero/Inizan et al_2023_Scalable hybrid deep neural networks-polarizable potentials biomolecular.pdf;/Users/wasmer/Zotero/storage/ZQFN36VU/Inizan et al. - 2023 - Scalable hybrid deep neural networkspolarizable p.pdf}
}
+@book{internationalenergyagencyWorldEnergyOutlook2022,
+ title = {World {{Energy Outlook}} 2022},
+ author = {{International Energy Agency}},
+ date = {2022-10-27},
+ series = {World {{Energy Outlook}}},
+ publisher = {OECD},
+ doi = {10.1787/3a469970-en},
+ url = {https://www.oecd-ilibrary.org/energy/world-energy-outlook-2022_3a469970-en},
+ urldate = {2024-08-02},
+ isbn = {978-92-64-42544-6},
+ langid = {english},
+ keywords = {/unread}
+}
+
@book{inuiGroupTheoryIts1990,
title = {Group {{Theory}} and {{Its Applications}} in {{Physics}}},
author = {Inui, Teturo and Tanabe, Yukito and Onodera, Yositaka},
@@ -7874,7 +8624,7 @@ Subject\_term\_id: computational-methods;research-management},
volume = {128},
number = {1-2},
eprint = {cond-mat/9909130},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
pages = {1--45},
issn = {00104655},
doi = {10.1016/S0010-4655(00)00072-2},
@@ -7916,7 +8666,7 @@ Subject\_term\_id: computational-methods;research-management},
volume = {2},
number = {5},
eprint = {2306.06283},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
pages = {1233--1250},
issn = {2635-098X},
@@ -7958,7 +8708,7 @@ Subject\_term\_id: computational-methods;research-management},
urldate = {2023-05-20},
abstract = {Machine learning has revolutionized many fields and has recently found applications in chemistry and materials science. The small datasets commonly found in chemistry lead to various sophisticated machine-learning approaches that incorporate chemical knowledge for each application and therefore require a lot of expertise to develop. Here, we show that large language models that have been trained on vast amounts of text extracted from the internet can easily be adapted to solve various tasks in chemistry and materials science by simply prompting them with chemical questions in natural language. We compared this approach with dedicated machine-learning models for many applications spanning properties of molecules and materials to the yield of chemical reactions. Surprisingly, we find this approach performs comparable to or even outperforms the conventional techniques, particularly in the low data limit. In addition, by simply inverting the questions, we can even perform inverse design successfully. The high performance, especially for small data sets, combined with the ease of use, can have a fundamental impact on how we leverage machine learning in the chemical and material sciences. Next to a literature search, querying a foundational model might become a routine way to bootstrap a project by leveraging the collective knowledge encoded in these foundational models.},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {alternative approaches,AML,chemistry,classification,generative models,GPT,GPT-3,high-entropy alloys,inverse design,IUPAC,library,LLM,ML,pretrained models,property prediction,regression,SELFIES,small data,SMILES,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Jablonka et al_2023_Is GPT-3 all you need for low-data discovery in chemistry.pdf}
}
@@ -8045,13 +8795,13 @@ Subject\_term\_id: computational-methods;research-management},
author = {Jain, Moksh and Deleu, Tristan and Hartford, Jason and Liu, Cheng-Hao and Hernandez-Garcia, Alex and Bengio, Yoshua},
date = {2023-02-01},
eprint = {2302.00615},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2302.00615},
url = {http://arxiv.org/abs/2302.00615},
urldate = {2023-06-23},
abstract = {Tackling the most pressing problems for humanity, such as the climate crisis and the threat of global pandemics, requires accelerating the pace of scientific discovery. While science has traditionally relied on trial and error and even serendipity to a large extent, the last few decades have seen a surge of data-driven scientific discoveries. However, in order to truly leverage large-scale data sets and high-throughput experimental setups, machine learning methods will need to be further improved and better integrated in the scientific discovery pipeline. A key challenge for current machine learning methods in this context is the efficient exploration of very large search spaces, which requires techniques for estimating reducible (epistemic) uncertainty and generating sets of diverse and informative experiments to perform. This motivated a new probabilistic machine learning framework called GFlowNets, which can be applied in the modeling, hypotheses generation and experimental design stages of the experimental science loop. GFlowNets learn to sample from a distribution given indirectly by a reward function corresponding to an unnormalized probability, which enables sampling diverse, high-reward candidates. GFlowNets can also be used to form efficient and amortized Bayesian posterior estimators for causal models conditioned on the already acquired experimental data. Having such posterior models can then provide estimators of epistemic uncertainty and information gain that can drive an experimental design policy. Altogether, here we will argue that GFlowNets can become a valuable tool for AI-driven scientific discovery, especially in scenarios of very large candidate spaces where we have access to cheap but inaccurate measurements or to expensive but accurate measurements. This is a common setting in the context of drug and material discovery, which we use as examples throughout the paper.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,Bayesian methods,drug discovery,General ML,generative models,materials discovery,MCMC,ML,molecular docking,molecules},
file = {/Users/wasmer/Nextcloud/Zotero/Jain et al_2023_GFlowNets for AI-Driven Scientific Discovery.pdf;/Users/wasmer/Zotero/storage/298HNEQA/2302.html}
}
@@ -8061,13 +8811,13 @@ Subject\_term\_id: computational-methods;research-management},
author = {Janakarajan, Nikita and Erdmann, Tim and Swaminathan, Sarath and Laino, Teodoro and Born, Jannis},
date = {2023-09-28},
eprint = {2309.16235},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, q-bio},
doi = {10.48550/arXiv.2309.16235},
url = {http://arxiv.org/abs/2309.16235},
urldate = {2023-10-05},
abstract = {The success of language models, especially transformer-based architectures, has trickled into other domains giving rise to "scientific language models" that operate on small molecules, proteins or polymers. In chemistry, language models contribute to accelerating the molecule discovery cycle as evidenced by promising recent findings in early-stage drug discovery. Here, we review the role of language models in molecular discovery, underlining their strength in de novo drug design, property prediction and reaction chemistry. We highlight valuable open-source software assets thus lowering the entry barrier to the field of scientific language modeling. Last, we sketch a vision for future molecular design that combines a chatbot interface with access to computational chemistry tools. Our contribution serves as a valuable resource for researchers, chemists, and AI enthusiasts interested in understanding how language models can and will be used to accelerate chemical discovery.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,biomolecules,chemical synthesis,chemistry,drug discovery,foundation models,IBM,LLM,ML,nlp,property prediction,review},
file = {/Users/wasmer/Nextcloud/Zotero/Janakarajan et al_2023_Language models in molecular discovery.pdf;/Users/wasmer/Zotero/storage/PR7IJC84/2309.html}
}
@@ -8077,13 +8827,13 @@ Subject\_term\_id: computational-methods;research-management},
author = {Janssen, Jan and Makarov, Edgar and Hickel, Tilmann and Shapeev, Alexander V. and Neugebauer, Jörg},
date = {2021-12-07},
eprint = {2112.04081},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2112.04081},
url = {http://arxiv.org/abs/2112.04081},
urldate = {2023-05-26},
abstract = {First principles approaches have revolutionized our ability in using computers to predict, explore and design materials. A major advantage commonly associated with these approaches is that they are fully parameter free. However, numerically solving the underlying equations requires to choose a set of convergence parameters. With the advent of high-throughput calculations it becomes exceedingly important to achieve a truly parameter free approach. Utilizing uncertainty quantification (UQ) and tensor decomposition we derive a numerically highly efficient representation of the statistical and systematic error in the multidimensional space of the convergence parameters. Based on this formalism we implement a fully automated approach that requires as input the target accuracy rather than convergence parameters. The performance and robustness of the approach are shown by applying it to a large set of elements crystallizing in a cubic fcc lattice.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,Condensed Matter - Materials Science},
file = {/Users/wasmer/Nextcloud/Zotero/Janssen et al_2021_Automated optimization of convergence parameters in plane wave density.pdf;/Users/wasmer/Zotero/storage/JGP5A47W/2112.html}
}
@@ -8107,6 +8857,22 @@ Subject\_term\_id: computational-methods;research-management},
file = {/Users/wasmer/Zotero/storage/TNV7XY35/S0927025618304786.html}
}
+@thesis{jaroszEfficientMonteCarlo2008,
+ type = {phdthesis},
+ title = {Efficient Monte Carlo Methods for Light Transport in Scattering Media},
+ author = {Jarosz, Wojciech},
+ date = {2008},
+ institution = {University of California at San Diego},
+ location = {USA},
+ url = {https://dl.acm.org/doi/10.5555/1559013},
+ abstract = {In this dissertation we focus on developing accurate and efficient Monte Carlo methods for synthesizing images containing general participating media. Participating media such as clouds, smoke, and fog are ubiquitous in the world and are responsible for many important visual phenomena which are of interest to computer graphics as well as related fields. When present, the medium participates in lighting interactions by scattering or absorbing photons as they travel through the scene. Though these effects add atmosphere and considerable depth to rendered images they are computationally very expensive to simulate. Most practical solutions make simplifying assumptions about the medium in order to maintain efficiency. Unfortunately, accurate and efficient simulation of light transport in general scattering media is a challenging undertaking. In this dissertation, we address this problem by introducing two complementary techniques. We first turn to the irradiance caching method for surface illumination. Irradiance caching gains efficiency by computing an accurate representation of lighting only at a sparse set of locations and reusing these values through interpolation whenever possible. We derive the mathematical concepts that form the foundation of this approach and analyze its strengths and weaknesses. Drawing inspiration from this algorithm, we then introduce a novel volumetric radiance caching method for efficiently simulating global illumination within participating media. In developing the technique we also introduce efficient methods for evaluating the gradient of the lighting within participating media. Our gradient analysis has immediate applicability for improved interpolation quality in both surface and media-based caching methods. We also develop a novel photon mapping technique for participating media. We present a theoretical reformulation of volumetric photon mapping, which provides significant new insights. This reformulation makes it easier to qualify the error introduced by the radiance estimate but, more importantly, also allows us to develop more efficient rendering techniques. Conventional photon mapping accelerate the computation of lighting at any point in the scene by performing density estimation. In contrast, our reformulation accelerates the computation of accumulated lighting along the length of entire rays . This algorithmic improvement provides for significantly reduced render times and even the potential for real-time visualization of light transport in participating media.},
+ pagetotal = {220},
+ keywords = {computer graphics,Computer science,educational,Jacobi-Legendre,mathematical tricks,mathematics,spherical harmonics,thesis},
+ annotation = {AAI3320228\\
+ISBN-13: 9780549720713},
+ file = {/Users/wasmer/Nextcloud/Zotero/Jarosz_2008_Efficient monte carlo methods for light transport in scattering media.pdf}
+}
+
@inproceedings{jiaPushingLimitMolecular2020,
title = {Pushing the {{Limit}} of {{Molecular Dynamics}} with {{Ab Initio Accuracy}} to 100 {{Million Atoms}} with {{Machine Learning}}},
booktitle = {{{SC20}}: {{International Conference}} for {{High Performance Computing}}, {{Networking}}, {{Storage}} and {{Analysis}}},
@@ -8228,14 +8994,14 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
author = {Jørgensen, Peter Bjørn and Bhowmik, Arghya},
date = {2020-11-04},
eprint = {2011.03346},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2011.03346},
url = {http://arxiv.org/abs/2011.03346},
urldate = {2022-07-10},
abstract = {We introduce DeepDFT, a deep learning model for predicting the electronic charge density around atoms, the fundamental variable in electronic structure simulations from which all ground state properties can be calculated. The model is formulated as neural message passing on a graph, consisting of interacting atom vertices and special query point vertices for which the charge density is predicted. The accuracy and scalability of the model are demonstrated for molecules, solids and liquids. The trained model achieves lower average prediction errors than the observed variations in charge density obtained from density functional theory simulations using different exchange correlation functionals.},
- pubstate = {preprint},
- keywords = {\_tablet,AML,DeepDFT,disordered,electrochemistry,equivariant,GNN,invariance,library,linear scaling,liquids,materials,ML,ML-DFT,ML-ESM,molecules,MPNN,NN,organic chemistry,original publication,oxides,PAiNN,prediction of electron density,PyTorch,QM9,SchNet,transition metals,VASP,with-code,with-data},
+ pubstate = {prepublished},
+ keywords = {AML,DeepDFT,disordered,electrochemistry,equivariant,GNN,invariance,library,linear scaling,liquids,materials,ML,ML-DFT,ML-ESM,molecules,MPNN,NN,organic chemistry,original publication,oxides,PAiNN,prediction of electron density,PyTorch,QM9,SchNet,transition metals,VASP,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Jørgensen_Bhowmik_2020_DeepDFT.pdf;/Users/wasmer/Zotero/storage/QXJKV745/2011.html}
}
@@ -8256,7 +9022,7 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
abstract = {Electron density \$\$\textbackslash rho (\textbackslash overrightarrow\{\{\{\{\textbackslash bf\{r\}\}\}\}\})\$\$is the fundamental variable in the calculation of ground state energy with density functional theory (DFT). Beyond total energy, features and changes in \$\$\textbackslash rho (\textbackslash overrightarrow\{\{\{\{\textbackslash bf\{r\}\}\}\}\})\$\$distributions are often used to capture critical physicochemical phenomena in functional materials. We present a machine learning framework for the prediction of \$\$\textbackslash rho (\textbackslash overrightarrow\{\{\{\{\textbackslash bf\{r\}\}\}\}\})\$\$. The model is based on equivariant graph neural networks and the electron density is predicted at special query point vertices that are part of the message-passing graph, but only receive messages. The model is tested across multiple datasets of molecules (QM9), liquid ethylene carbonate electrolyte (EC) and LixNiyMnzCo(1-y-z)O2 lithium ion battery cathodes (NMC). For QM9 molecules, the accuracy of the proposed model exceeds typical variability in \$\$\textbackslash rho (\textbackslash overrightarrow\{\{\{\{\textbackslash bf\{r\}\}\}\}\})\$\$obtained from DFT done with different exchange-correlation functionals. The accuracy on all three datasets is beyond state of the art and the computation time is orders of magnitude faster than DFT.},
issue = {1},
langid = {english},
- keywords = {\_tablet,AML,DeepDFT,disordered,electrochemistry,equivariant,GNN,invariance,library,linear scaling,liquids,materials,ML,ML-DFT,ML-ESM,molecules,MPNN,NN,organic chemistry,original publication,oxides,PAiNN,prediction of electron density,PyTorch,QM9,SchNet,transition metals,VASP,with-code,with-data},
+ keywords = {AML,DeepDFT,disordered,electrochemistry,equivariant,GNN,invariance,library,linear scaling,liquids,materials,ML,ML-DFT,ML-ESM,molecules,MPNN,NN,organic chemistry,original publication,oxides,PAiNN,prediction of electron density,PyTorch,QM9,SchNet,transition metals,VASP,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Jørgensen_Bhowmik_2022_Equivariant graph neural networks for fast electron density estimation of.pdf}
}
@@ -8265,14 +9031,14 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
author = {Jørgensen, Peter Bjørn and Bhowmik, Arghya},
date = {2021-12-01},
eprint = {2112.00652},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2112.00652},
url = {http://arxiv.org/abs/2112.00652},
urldate = {2022-07-10},
abstract = {Electron density \$\textbackslash rho(\textbackslash vec\{r\})\$ is the fundamental variable in the calculation of ground state energy with density functional theory (DFT). Beyond total energy, features in \$\textbackslash rho(\textbackslash vec\{r\})\$ distribution and modifications in \$\textbackslash rho(\textbackslash vec\{r\})\$ are often used to capture critical physicochemical phenomena in functional materials and molecules at the electronic scale. Methods providing access to \$\textbackslash rho(\textbackslash vec\{r\})\$ of complex disordered systems with little computational cost can be a game changer in the expedited exploration of materials phase space towards the inverse design of new materials with better functionalities. We present a machine learning framework for the prediction of \$\textbackslash rho(\textbackslash vec\{r\})\$. The model is based on equivariant graph neural networks and the electron density is predicted at special query point vertices that are part of the message passing graph, but only receive messages. The model is tested across multiple data sets of molecules (QM9), liquid ethylene carbonate electrolyte (EC) and LixNiyMnzCo(1-y-z)O2 lithium ion battery cathodes (NMC). For QM9 molecules, the accuracy of the proposed model exceeds typical variability in \$\textbackslash rho(\textbackslash vec\{r\})\$ obtained from DFT done with different exchange-correlation functional and show beyond the state of the art accuracy. The accuracy is even better for the mixed oxide (NMC) and electrolyte (EC) datasets. The linear scaling model's capacity to probe thousands of points simultaneously permits calculation of \$\textbackslash rho(\textbackslash vec\{r\})\$ for large complex systems many orders of magnitude faster than DFT allowing screening of disordered functional materials.},
- pubstate = {preprint},
- keywords = {\_tablet,AML,DeepDFT,disordered,electrochemistry,equivariant,GNN,invariance,library,linear scaling,liquids,materials,ML,ML-DFT,ML-ESM,molecules,MPNN,NN,organic chemistry,original publication,oxides,PAiNN,prediction of electron density,PyTorch,QM9,SchNet,transition metals,VASP,with-code,with-data},
+ pubstate = {prepublished},
+ keywords = {AML,DeepDFT,disordered,electrochemistry,equivariant,GNN,invariance,library,linear scaling,liquids,materials,ML,ML-DFT,ML-ESM,molecules,MPNN,NN,organic chemistry,original publication,oxides,PAiNN,prediction of electron density,PyTorch,QM9,SchNet,transition metals,VASP,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Jørgensen_Bhowmik_2021_Graph neural networks for fast electron density estimation of molecules,.pdf;/Users/wasmer/Zotero/storage/MBXG22TT/2112.html}
}
@@ -8281,13 +9047,13 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
author = {Joshi, Chaitanya K. and Bodnar, Cristian and Mathis, Simon V. and Cohen, Taco and Liò, Pietro},
date = {2023-06-03},
eprint = {2301.09308},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math, stat},
doi = {10.48550/arXiv.2301.09308},
url = {http://arxiv.org/abs/2301.09308},
urldate = {2023-10-07},
abstract = {The expressive power of Graph Neural Networks (GNNs) has been studied extensively through the Weisfeiler-Leman (WL) graph isomorphism test. However, standard GNNs and the WL framework are inapplicable for geometric graphs embedded in Euclidean space, such as biomolecules, materials, and other physical systems. In this work, we propose a geometric version of the WL test (GWL) for discriminating geometric graphs while respecting the underlying physical symmetries: permutations, rotation, reflection, and translation. We use GWL to characterise the expressive power of geometric GNNs that are invariant or equivariant to physical symmetries in terms of distinguishing geometric graphs. GWL unpacks how key design choices influence geometric GNN expressivity: (1) Invariant layers have limited expressivity as they cannot distinguish one-hop identical geometric graphs; (2) Equivariant layers distinguish a larger class of graphs by propagating geometric information beyond local neighbourhoods; (3) Higher order tensors and scalarisation enable maximally powerful geometric GNNs; and (4) GWL's discrimination-based perspective is equivalent to universal approximation. Synthetic experiments supplementing our results are available at \textbackslash url\{https://github.com/chaitjo/geometric-gnn-dojo\}},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,benchmarking,DimeNet,educational,equivariant,General ML,geometric deep learning,GNN,invariance,library,MACE,ML,MPNN,review,review-of-GNN,SchNet,SE(3),transformer,with-code,WL test},
file = {/Users/wasmer/Nextcloud/Zotero/Joshi et al_2023_On the Expressive Power of Geometric Graph Neural Networks.pdf;/Users/wasmer/Zotero/storage/4HKE6SLD/2301.html}
}
@@ -8322,13 +9088,13 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
author = {Kaba, Sékou-Oumar and Ravanbakhsh, Siamak},
date = {2023-01-15},
eprint = {2211.15420},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2211.15420},
url = {http://arxiv.org/abs/2211.15420},
urldate = {2023-04-03},
abstract = {Supervised learning with deep models has tremendous potential for applications in materials science. Recently, graph neural networks have been used in this context, drawing direct inspiration from models for molecules. However, materials are typically much more structured than molecules, which is a feature that these models do not leverage. In this work, we introduce a class of models that are equivariant with respect to crystalline symmetry groups. We do this by defining a generalization of the message passing operations that can be used with more general permutation groups, or that can alternatively be seen as defining an expressive convolution operation on the crystal graph. Empirically, these models achieve competitive results with state-of-the-art on property prediction tasks.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,CGCNN,crystal structure,crystal symmetry,ECN,equivariant,GNN,group theory,magnetic moment,materials,materials project,MEGNet,ML,ML-DFT,MPNN,original publication,prediction of energy,prediction of magnetic moment,supercell,voronoi tessellation},
file = {/Users/wasmer/Nextcloud/Zotero/Kaba_Ravanbakhsh_2023_Equivariant Networks for Crystal Structures.pdf;/Users/wasmer/Zotero/storage/2YADPZ3J/2211.html}
}
@@ -8347,7 +9113,7 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
url = {https://link.aps.org/doi/10.1103/PhysRevMaterials.7.044407},
urldate = {2023-06-01},
abstract = {Magnetic materials are crucial components of many technologies that could drive the ecological transition, including electric motors, wind turbine generators, and magnetic refrigeration systems. Discovering materials with large magnetic moments is therefore an increasing priority. Here, using state-of-the-art machine learning methods, we scan the Inorganic Crystal Structure Database (ICSD) of hundreds of thousands of existing materials to find those that are ferromagnetic and have large magnetic moments. Crystal graph convolutional neural networks (CGCNNs), materials graph network (MEGNet), and random forests are trained on the Materials Project database that contains the results of high-throughput density functional theory (DFT) predictions. For random forests, we use a stochastic method to select nearly 100 relevant descriptors based on chemical composition and crystal structure. This gives results that are comparable to those of neural networks. Our findings suggests that magnetic properties are intrinsically more difficult to predict than other DFT-calculated properties. The comparison between the different machine learning approaches gives an estimate of the errors for our predictions on the ICSD database. Validating our final predictions by comparisons with available experimental data, we found 15 materials that are likely to have large magnetic moments and have not yet been studied experimentally.},
- keywords = {\_tablet,AML,benchmarking,CGCNN,compositional descriptors,descriptors,DFT,Ferromagnetism,GGA,GGA+U,GNN,HTC,ICSD,magnetic moment,magnetism,magnetocaloric,materials,materials project,materials screening,MEGNet,ML,model comparison,prediction from structure,prediction of magnetic moment,random forest,rare earths,transition metals},
+ keywords = {AML,benchmarking,CGCNN,compositional descriptors,descriptors,DFT,Ferromagnetism,GGA,GGA+U,GNN,HTC,ICSD,magnetic moment,magnetism,magnetocaloric,materials,materials project,materials screening,MEGNet,ML,model comparison,prediction from structure,prediction of magnetic moment,random forest,rare earths,transition metals},
file = {/Users/wasmer/Nextcloud/Zotero/Kaba et al_2023_Prediction of large magnetic moment materials with graph neural networks and.pdf;/Users/wasmer/Zotero/storage/ES8ANMU8/PhysRevMaterials.7.html}
}
@@ -8414,7 +9180,7 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
author = {Kalita, Bhupalee and Burke, Kieron},
date = {2021-12-03},
eprint = {2112.05554},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
url = {http://arxiv.org/abs/2112.05554},
urldate = {2022-03-28},
@@ -8475,6 +9241,26 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
file = {/Users/wasmer/Nextcloud/Zotero/Kanter_Veeramachaneni_2015_Deep feature synthesis.pdf;/Users/wasmer/Zotero/storage/FTYFE5ZI/7344858.html}
}
+@article{karniadakisPhysicsinformedMachineLearning2021,
+ title = {Physics-Informed Machine Learning},
+ author = {Karniadakis, George Em and Kevrekidis, Ioannis G. and Lu, Lu and Perdikaris, Paris and Wang, Sifan and Yang, Liu},
+ date = {2021-06},
+ journaltitle = {Nature Reviews Physics},
+ shortjournal = {Nat Rev Phys},
+ volume = {3},
+ number = {6},
+ pages = {422--440},
+ publisher = {Nature Publishing Group},
+ issn = {2522-5820},
+ doi = {10.1038/s42254-021-00314-5},
+ url = {https://www.nature.com/articles/s42254-021-00314-5},
+ urldate = {2024-06-14},
+ abstract = {Despite great progress in simulating multiphysics problems using the numerical discretization of partial differential equations (PDEs), one still cannot seamlessly incorporate noisy data into existing algorithms, mesh generation remains complex, and high-dimensional problems governed by parameterized PDEs cannot be tackled. Moreover, solving inverse problems with hidden physics is often prohibitively expensive and requires different formulations and elaborate computer codes. Machine learning has emerged as a promising alternative, but training deep neural networks requires big data, not always available for scientific problems. Instead, such networks can be trained from additional information obtained by enforcing the physical laws (for example, at random points in the continuous space-time domain). Such physics-informed learning integrates (noisy) data and mathematical models, and implements them through neural networks or other kernel-based regression networks. Moreover, it may be possible to design specialized network architectures that automatically satisfy some of the physical invariants for better accuracy, faster training and improved generalization. Here, we review some of the prevailing trends in embedding physics into machine learning, present some of the current capabilities and limitations and discuss diverse applications of physics-informed learning both for forward and inverse problems, including discovering hidden physics and tackling high-dimensional problems.},
+ langid = {english},
+ keywords = {/unread,AI4Science,AML,DeePMD-kit,DeepONet,FermiNet,ML,ML-PDE,MLP,neural operator,physics-informed ML,PINN,review,review-of-AI4science,review-of-PIML,scaling},
+ file = {/Users/wasmer/Nextcloud/Zotero/Karniadakis et al_2021_Physics-informed machine learning.pdf}
+}
+
@article{kasimLearningExchangecorrelationFunctional2021,
title = {Learning the Exchange-Correlation Functional from Nature with Fully Differentiable Density Functional Theory},
author = {Kasim, Muhammad F. and Vinko, Sam M.},
@@ -8484,7 +9270,7 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
volume = {127},
number = {12},
eprint = {2102.04229},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
pages = {126403},
issn = {0031-9007, 1079-7114},
doi = {10.1103/PhysRevLett.127.126403},
@@ -8510,7 +9296,7 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
urldate = {2022-06-30},
abstract = {Machine learning (ML)-based models have greatly enhanced the traditional materials discovery and design pipeline. Specifically, in recent years, surrogate ML models for material property prediction have demonstrated success in predicting discrete scalar-valued target properties to within reasonable accuracy of their DFT-computed values. However, accurate prediction of spectral targets, such as the electron density of states (DOS), poses a much more challenging problem due to the complexity of the target, and the limited amount of available training data. In this study, we present an extension of the recently developed atomistic line graph neural network to accurately predict DOS of a large set of material unit cell structures, trained to the publicly available JARVIS-DFT dataset. Furthermore, we evaluate two methods of representation of the target quantity: a direct discretized spectrum, and a compressed low-dimensional representation obtained using an autoencoder. Through this work, we demonstrate the utility of graph-based featurization and modeling methods in the prediction of complex targets that depend on both chemistry and directional characteristics of material structures.},
langid = {english},
- keywords = {\_tablet,ALIGNN,autoencoder,DFT,dimensionality reduction,dimensionality reduction of target,GNN,JARVIS,JARVIS-DFT,ML,ML-DFT,ML-ESM,MPNN,prediction from structure,prediction of LDOS},
+ keywords = {ALIGNN,autoencoder,DFT,dimensionality reduction,dimensionality reduction of target,GNN,JARVIS,JARVIS-DFT,ML,ML-DFT,ML-ESM,MPNN,prediction from structure,prediction of LDOS},
file = {/Users/wasmer/Nextcloud/Zotero/Kaundinya et al_2022_Prediction of the Electron Density of States for Crystalline Compounds with.pdf}
}
@@ -8527,7 +9313,7 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
url = {https://pubs.rsc.org/en/content/articlelanding/2023/dd/d2dd00133k},
urldate = {2023-08-19},
langid = {english},
- keywords = {Citrine Informatics,todo-tagging},
+ keywords = {Citrine Informatics,hybrid AI/simulation,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Kavalsky et al_2023_By how much can closed-loop frameworks accelerate computational materials.pdf}
}
@@ -8582,7 +9368,7 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
abstract = {Two-dimensional materials offer a promising platform for the next generation of (opto-) electronic devices and other high technology applications. One of the most exciting characteristics of 2D crystals is the ability to tune their properties via controllable introduction of defects. However, the search space for such structures is enormous, and ab-initio computations prohibitively expensive. We propose a machine learning approach for rapid estimation of the properties of 2D material given the lattice structure and defect configuration. The method suggests a way to represent configuration of 2D materials with defects that allows a neural network to train quickly and accurately. We compare our methodology with the state-of-the-art approaches and demonstrate at least 3.7 times energy prediction error drop. Also, our approach is an order of magnitude more resource-efficient than its contenders both for the training and inference part.},
issue = {1},
langid = {english},
- keywords = {\_tablet,2D material,AML,defects,disordered,exchange interaction,GNN,library,ML,multi-defect,point defects,prediction from defect structure,prediction of formation energy,prediction of HOMO/LUMO,representation of defects,sparsification,with-code,with-data,with-demo},
+ keywords = {2D material,AML,defects,disordered,exchange interaction,GNN,library,ML,multi-defect,point defects,prediction from defect structure,prediction of formation energy,prediction of HOMO/LUMO,representation of defects,sparsification,with-code,with-data,with-demo},
file = {/Users/wasmer/Nextcloud/Zotero/Kazeev et al_2023_Sparse representation for machine learning the properties of defects in 2D.pdf}
}
@@ -8623,6 +9409,7 @@ Subject\_term: Condensed-matter physics, Physics, Materials science},
abstract = {The physical description of all materials is rooted in quantum mechanics, which describes how atoms bond and electrons interact at a fundamental level. Although these quantum effects can in many cases be approximated by a classical description at the macroscopic level, in recent years there has been growing interest in material systems where quantum effects remain manifest over a wider range of energy and length scales. Such quantum materials include superconductors, graphene, topological insulators, Weyl semimetals, quantum spin liquids, and spin ices. Many of them derive their properties from reduced dimensionality, in particular from confinement of electrons to two-dimensional sheets. Moreover, they tend to be materials in which electrons cannot be considered as independent particles but interact strongly and give rise to collective excitations known as quasiparticles. In all cases, however, quantum-mechanical effects fundamentally alter properties of the material. This Review surveys the electronic properties of quantum materials through the prism of the electron wavefunction, and examines how its entanglement and topology give rise to a rich variety of quantum states and phases; these are less classically describable than conventional ordered states also driven by quantum mechanics, such as ferromagnetism.},
issue = {11},
langid = {english},
+ keywords = {/unread,for introductions,Hall effect,Hall FQHE,Hall QHE,High-Tc Cuprates,Kitaev magnet,magnetism,Majorana,MZM,physics,quantum materials,review,review-of-TIs,semimetal,SOC,superconductor,topological,topological insulator,Topological Superconductor},
annotation = {Bandiera\_abtest: a\\
Cg\_type: Nature Research Journals\\
Primary\_atype: Reviews\\
@@ -8674,17 +9461,33 @@ Subject\_term\_id: quantum-physics;theoretical-physics},
author = {Kidger, Patrick},
date = {2022-02-04},
eprint = {2202.02435},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math, stat},
doi = {10.48550/arXiv.2202.02435},
url = {http://arxiv.org/abs/2202.02435},
urldate = {2022-09-07},
abstract = {The conjoining of dynamical systems and deep learning has become a topic of great interest. In particular, neural differential equations (NDEs) demonstrate that neural networks and differential equation are two sides of the same coin. Traditional parameterised differential equations are a special case. Many popular neural network architectures, such as residual networks and recurrent networks, are discretisations. NDEs are suitable for tackling generative problems, dynamical systems, and time series (particularly in physics, finance, ...) and are thus of interest to both modern machine learning and traditional mathematical modelling. NDEs offer high-capacity function approximation, strong priors on model space, the ability to handle irregular data, memory efficiency, and a wealth of available theory on both sides. This doctoral thesis provides an in-depth survey of the field. Topics include: neural ordinary differential equations (e.g. for hybrid neural/mechanistic modelling of physical systems); neural controlled differential equations (e.g. for learning functions of irregular time series); and neural stochastic differential equations (e.g. to produce generative models capable of representing complex stochastic dynamics, or sampling from complex high-dimensional distributions). Further topics include: numerical methods for NDEs (e.g. reversible differential equations solvers, backpropagation through differential equations, Brownian reconstruction); symbolic regression for dynamical systems (e.g. via regularised evolution); and deep implicit models (e.g. deep equilibrium models, differentiable optimisation). We anticipate this thesis will be of interest to anyone interested in the marriage of deep learning with dynamical systems, and hope it will provide a useful reference for the current state of the art.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {backpropagation,Deep learning,differential equations,NDE,PINN,thesis},
file = {/Users/wasmer/Nextcloud/Zotero/Kidger_2022_On Neural Differential Equations.pdf;/Users/wasmer/Zotero/storage/EHARV7VZ/2202.html}
}
+@online{kimGaussianPlaneWaveNeural2024,
+ title = {Gaussian {{Plane-Wave Neural Operator}} for {{Electron Density Estimation}}},
+ author = {Kim, Seongsu and Ahn, Sungsoo},
+ date = {2024-02-05},
+ eprint = {2402.04278},
+ eprinttype = {arXiv},
+ eprintclass = {physics},
+ doi = {10.48550/arXiv.2402.04278},
+ url = {http://arxiv.org/abs/2402.04278},
+ urldate = {2024-05-27},
+ abstract = {This work studies machine learning for electron density prediction, which is fundamental for understanding chemical systems and density functional theory (DFT) simulations. To this end, we introduce the Gaussian plane-wave neural operator (GPWNO), which operates in the infinite-dimensional functional space using the plane-wave and Gaussian-type orbital bases, widely recognized in the context of DFT. In particular, both high- and low-frequency components of the density can be effectively represented due to the complementary nature of the two bases. Extensive experiments on QM9, MD, and material project datasets demonstrate GPWNO's superior performance over ten baselines.},
+ pubstate = {prepublished},
+ keywords = {AML,DFT,equivariant,FNO,graphon convolution,materials,materials project,MD17,ML,ML-Density,ML-DFT,ML-ESM,molecules,neural operator,plane-wave,prediction of electron density,QM9},
+ file = {/Users/wasmer/Nextcloud/Zotero/Kim_Ahn_2024_Gaussian Plane-Wave Neural Operator for Electron Density Estimation.pdf;/Users/wasmer/Zotero/storage/U6I47ZEN/2402.html}
+}
+
@article{kimMachinelearnedMetricsPredicting2020,
title = {Machine-Learned Metrics for Predicting the Likelihood of Success in Materials Discovery},
author = {Kim, Yoolhee and Kim, Edward and Antono, Erin and Meredig, Bryce and Ling, Julia},
@@ -8712,13 +9515,13 @@ Subject\_term\_id: quantum-physics;theoretical-physics},
author = {Kingma, Diederik P. and Ba, Jimmy},
date = {2017-01-29},
eprint = {1412.6980},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.1412.6980},
url = {http://arxiv.org/abs/1412.6980},
urldate = {2023-07-21},
abstract = {We introduce Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments. The method is straightforward to implement, is computationally efficient, has little memory requirements, is invariant to diagonal rescaling of the gradients, and is well suited for problems that are large in terms of data and/or parameters. The method is also appropriate for non-stationary objectives and problems with very noisy and/or sparse gradients. The hyper-parameters have intuitive interpretations and typically require little tuning. Some connections to related algorithms, on which Adam was inspired, are discussed. We also analyze the theoretical convergence properties of the algorithm and provide a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework. Empirical results demonstrate that Adam works well in practice and compares favorably to other stochastic optimization methods. Finally, we discuss AdaMax, a variant of Adam based on the infinity norm.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Adam,Deep learning,General ML,ML,NN,optimization,optimizer,original publication},
file = {/Users/wasmer/Nextcloud/Zotero/Kingma_Ba_2017_Adam.pdf;/Users/wasmer/Zotero/storage/FDMFEJS4/1412.html}
}
@@ -8753,8 +9556,8 @@ Subject\_term\_id: quantum-physics;theoretical-physics},
urldate = {2024-01-12},
abstract = {The anomalous Hall effect has been front and center in solid state research and material science for over a century now, and the complex transport phenomena in nontrivial magnetic textures have gained an increasing amount of attention, both in theoretical and experimental studies. However, a clear path forward to capturing the influence of magnetization dynamics on anomalous Hall effect even in smallest frustrated magnets or spatially extended magnetic textures is still intensively sought after. In this work, we present an expansion of the anomalous Hall tensor into symmetrically invariant objects, encoding the magnetic configuration up to arbitrary power of spin. We show that these symmetric invariants can be utilized in conjunction with advanced regularization techniques in order to build models for the electric transport in magnetic textures which are, on one hand, complete with respect to the point group symmetry of the underlying lattice, and on the other hand, depend on a minimal number of order parameters only. Here, using a four-band tight-binding model on a honeycomb lattice, we demonstrate that the developed method can be used to address the importance and properties of higher-order contributions to transverse transport. The efficiency and breadth enabled by this method provides an ideal systematic approach to tackle the inherent complexity of response properties of noncollinear magnets, paving the way to the exploration of electric transport in intrinsically frustrated magnets as well as large-scale magnetic textures.},
langid = {english},
- pubstate = {preprint},
- keywords = {2D,2D material,AML,electric transport,feature selection,group theory,Hall AHE,Hall effect,higher order,invariance,linear regression,magnetic structure,magnetic supperlattice,magnetism,materials,ML,non-collinear,PCA,physics,point group,spin invariant,spin-dependent,spintronics,SVD,symmetrization,symmetry,TB,tensor decomposition,tight binding},
+ pubstate = {prepublished},
+ keywords = {2D,2D material,AML,electric transport,feature selection,group theory,Hall AHE,Hall effect,higher order,honeycomb lattice,invariance,linear regression,magnetic structure,magnetic supperlattice,magnetism,materials,ML,non-collinear,PCA,physics,point group,spin invariant,spin-dependent,spintronics,SVD,symmetrization,symmetry,TB,tensor decomposition,tight binding},
file = {/Users/wasmer/Nextcloud/Zotero/Kipp et al_2024_Machine learning inspired models for Hall effects in non-collinear magnets.pdf}
}
@@ -8780,7 +9583,7 @@ Subject\_term\_id: quantum-physics;theoretical-physics},
author = {Klicpera, Johannes and Becker, Florian and Günnemann, Stephan},
date = {2021-11-29},
eprint = {2106.08903},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
url = {http://arxiv.org/abs/2106.08903},
urldate = {2022-01-02},
@@ -8814,7 +9617,7 @@ Subject\_term\_id: quantum-physics;theoretical-physics},
author = {Klus, Stefan and Gelß, Patrick and Nüske, Feliks and Noé, Frank},
date = {2021-03-31},
eprint = {2103.17233},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {math-ph, physics:physics, physics:quant-ph, stat},
url = {http://arxiv.org/abs/2103.17233},
urldate = {2021-05-13},
@@ -8847,7 +9650,7 @@ Subject\_term\_id: quantum-physics;theoretical-physics},
author = {Knøsgaard, Nikolaj Rørbæk and Thygesen, Kristian Sommer},
date = {2021-07-13},
eprint = {2107.06029},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
url = {http://arxiv.org/abs/2107.06029},
urldate = {2021-08-05},
@@ -8886,6 +9689,23 @@ Subject\_term\_id: quantum-physics;theoretical-physics},
keywords = {/unread,educational,Julia,learning material,mathematics,online book,optimization,textbook,with-code}
}
+@article{kochkovNeuralGeneralCirculation2024,
+ title = {Neural General Circulation Models for Weather and Climate},
+ author = {Kochkov, Dmitrii and Yuval, Janni and Langmore, Ian and Norgaard, Peter and Smith, Jamie and Mooers, Griffin and Klöwer, Milan and Lottes, James and Rasp, Stephan and Düben, Peter and Hatfield, Sam and Battaglia, Peter and Sanchez-Gonzalez, Alvaro and Willson, Matthew and Brenner, Michael P. and Hoyer, Stephan},
+ date = {2024-07-22},
+ journaltitle = {Nature},
+ pages = {1--7},
+ publisher = {Nature Publishing Group},
+ issn = {1476-4687},
+ doi = {10.1038/s41586-024-07744-y},
+ url = {https://www.nature.com/articles/s41586-024-07744-y},
+ urldate = {2024-08-02},
+ abstract = {General circulation models (GCMs) are the foundation of weather and climate prediction1,2. GCMs are physics-based simulators that combine a numerical solver for large-scale dynamics with tuned representations for small-scale processes such as cloud formation. Recently, machine-learning models trained on reanalysis data have achieved comparable or better skill than GCMs for deterministic weather forecasting3,4. However, these models have not demonstrated improved ensemble forecasts, or shown sufficient stability for long-term weather and climate simulations. Here we present a GCM that combines a differentiable solver for atmospheric dynamics with machine-learning components and show that it can generate forecasts of deterministic weather, ensemble weather and climate on par with the best machine-learning and physics-based methods. NeuralGCM is competitive with machine-learning models for one- to ten-day forecasts, and with the European Centre for Medium-Range Weather Forecasts ensemble prediction for one- to fifteen-day forecasts. With prescribed sea surface temperature, NeuralGCM can accurately track climate metrics for multiple decades, and climate forecasts with 140-kilometre resolution show emergent phenomena such as realistic frequency and trajectories of tropical cyclones. For both weather and climate, our approach offers orders of magnitude computational savings over conventional GCMs, although our model does not extrapolate to substantially different future climates. Our results show that end-to-end deep learning is compatible with tasks performed by conventional GCMs and can enhance the large-scale physical simulations that are essential for understanding and predicting the Earth system.},
+ langid = {english},
+ keywords = {/unread,AI4Science,Google,hybrid AI/simulation,large models,PDE,weather forecasting},
+ file = {/Users/wasmer/Nextcloud/Zotero/Kochkov et al_2024_Neural general circulation models for weather and climate.pdf}
+}
+
@article{koFourthgenerationHighdimensionalNeural2021,
title = {A Fourth-Generation High-Dimensional Neural Network Potential with Accurate Electrostatics Including Non-Local Charge Transfer},
author = {Ko, Tsz Wai and Finkler, Jonas A. and Goedecker, Stefan and Behler, Jörg},
@@ -9025,6 +9845,22 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
file = {/Users/wasmer/Nextcloud/Zotero/Kohn_Rostoker_1954_Solution of the Schr-odinger Equation in Periodic Lattices with an Application.pdf;/Users/wasmer/Zotero/storage/RDPKHK8A/PhysRev.94.html}
}
+@online{kokerHigherOrderEquivariantNeural2024,
+ title = {Higher-{{Order Equivariant Neural Networks}} for {{Charge Density Prediction}} in {{Materials}}},
+ author = {Koker, Teddy and Quigley, Keegan and Taw, Eric and Tibbetts, Kevin and Li, Lin},
+ date = {2024-05-14},
+ eprint = {2312.05388},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat, physics:physics},
+ doi = {10.48550/arXiv.2312.05388},
+ url = {http://arxiv.org/abs/2312.05388},
+ urldate = {2024-05-27},
+ abstract = {The calculation of electron density distribution using density functional theory (DFT) in materials and molecules is central to the study of their quantum and macro-scale properties, yet accurate and efficient calculation remains a long-standing challenge. We introduce ChargE3Net, an E(3)-equivariant graph neural network for predicting electron density in atomic systems. ChargE3Net enables the learning of higher-order equivariant feature to achieve high predictive accuracy and model expressivity. We show that ChargE3Net exceeds the performance of prior work on diverse sets of molecules and materials. When trained on the massive dataset of over 100K materials in the Materials Project database, our model is able to capture the complexity and variability in the data, leading to a significant 26.7\% reduction in self-consistent iterations when used to initialize DFT calculations on unseen materials. Furthermore, we show that non-self-consistent DFT calculations using our predicted charge densities yield near-DFT performance on electronic and thermodynamic property prediction at a fraction of the computational cost. Further analysis attributes the greater predictive accuracy to improved modeling of systems with high angular variations. These results illuminate a pathway towards a machine learning-accelerated ab initio calculations for materials discovery.},
+ pubstate = {prepublished},
+ keywords = {AML,e3nn,equivariant,GNoME,hybrid AI/simulation,initial guess,large dataset,linear-scaling DFT,materials,materials project,ML,ML-Density,ML-DFT,ML-ESM,molecules,oxides,prediction of electron density,QM9,SCF,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Koker et al_2024_Higher-Order Equivariant Neural Networks for Charge Density Prediction in.pdf;/Users/wasmer/Zotero/storage/EN3MDLFU/2312.html}
+}
+
@article{kolbDiscoveringChargeDensity2017,
title = {Discovering Charge Density Functionals and Structure-Property Relationships with {{PROPhet}}: {{A}} General Framework for Coupling Machine Learning and First-Principles Methods},
shorttitle = {Discovering Charge Density Functionals and Structure-Property Relationships with {{PROPhet}}},
@@ -9153,8 +9989,8 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
urldate = {2023-04-10},
abstract = {Machine learning applications often need large amounts of training data to perform well. Whereas unlabeled data can be easily gathered, the labeling process is difficult, time-consuming, or expensive in most applications. Active learning can help solve this problem by querying labels for those data points that will improve the performance the most. Thereby, the goal is that the learning algorithm performs sufficiently well with fewer labels. We provide a library called scikit-activeml that covers the most relevant query strategies and implements tools to work with partially labeled data. It is programmed in Python and builds on top of scikit-learn.},
langid = {english},
- pubstate = {preprint},
- keywords = {/unread,\_tablet,active learning,General ML,library,Python,scikit-learn},
+ pubstate = {prepublished},
+ keywords = {/unread,active learning,General ML,library,Python,scikit-learn},
file = {/Users/wasmer/Nextcloud/Zotero/Kottke et al_2021_scikit-activeml.pdf}
}
@@ -9172,27 +10008,44 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
urldate = {2023-04-06},
abstract = {When the magnetic order is introduced into topological insulators (TIs), the time-reversal symmetry (TRS) is broken, and the non-trivial topological surface is driven into a new massive Dirac fermions state. The study of such TRS-breaking systems is one of the most emerging frontiers in condensed-matter physics. In this review, we outline the methods to break the TRS of the topological surface states. With robust out-of-plane magnetic order formed, we describe the intrinsic magnetisms in the magnetically doped 3D TI materials and the approach to manipulate each contribution. Most importantly, we summarize the theoretical developments and experimental observations of the scale-invariant quantum anomalous Hall effect (QAHE) in both the 2D and 3D Cr-doped (BiSb)2Te3 systems; at the same time, we also discuss the correlations between QAHE and other quantum transport phenomena. Finally, we highlight the use of TI/Cr-doped TI heterostructures to both manipulate the surface-related ferromagnetism and realize electrical manipulation of magnetization through the giant spin–orbit torques.},
langid = {english},
- keywords = {\_tablet,breaking of TRS,defects,Hall effect,Hall QAHE,heterostructures,magnetic doping,magnetic heterostructures,magnetism,physics,review,SOC,spin-dependent,Spin-orbit effects,topological insulator,TRS},
+ keywords = {breaking of TRS,defects,Hall effect,Hall QAHE,heterostructures,magnetic doping,magnetic heterostructures,magnetism,physics,review,SOC,spin-dependent,Spin-orbit effects,topological insulator,TRS},
file = {/Users/wasmer/Nextcloud/Zotero/Kou et al_2015_Magnetic topological insulators and quantum anomalous hall effect.pdf;/Users/wasmer/Zotero/storage/IMCZD9AZ/S0038109814004438.html}
}
@online{kovacsEvaluationMACEForce2023,
title = {Evaluation of the {{MACE Force Field Architecture}}: From {{Medicinal Chemistry}} to {{Materials Science}}},
shorttitle = {Evaluation of the {{MACE Force Field Architecture}}},
- author = {Kovacs, David Peter and Batatia, Ilyes and Arany, Eszter Sara and Csanyi, Gabor},
+ author = {Kovacs, David Peter and Batatia, Ilyes and Arany, Eszter Sara and Csányi, Gábor},
date = {2023-05-23},
eprint = {2305.14247},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2305.14247},
- url = {http://arxiv.org/abs/2305.14247},
+ url = {http://arxiv.org/abs/2305.14247v1},
urldate = {2023-05-26},
abstract = {The MACE architecture represents the state of the art in the field of machine learning force fields for a variety of in-domain, extrapolation and low-data regime tasks. In this paper, we further evaluate MACE by fitting models for published benchmark datasets. We show that MACE generally outperforms alternatives for a wide range of systems from amorphous carbon and general small molecule organic chemistry to large molecules and liquid water. We demonstrate the capabilities of the model on tasks ranging from constrained geometry optimisation to molecular dynamics simulations and find excellent performance across all tested domains. We show that MACE is very data efficient, and can reproduce experimental molecular vibrational spectra when trained on as few as 50 randomly selected reference configurations. We further demonstrate that the strictly local atom-centered model is sufficient for such tasks even in the case of large molecules and weakly interacting molecular assemblies.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ACE,AML,benchmarking,carbon,disordered,GNN,M3GNet,MACE,materials,ML,ML-FF,MLP,model comparison,model evaluation,molecules,MPNN,NewtonNet,QM9},
file = {/Users/wasmer/Nextcloud/Zotero/Kovacs et al_2023_Evaluation of the MACE Force Field Architecture.pdf;/Users/wasmer/Zotero/storage/NJT7SCKS/2305.html}
}
+@online{kovacsEvaluationMACEForce2023a,
+ title = {Evaluation of the {{MACE Force Field Architecture}}: From {{Medicinal Chemistry}} to {{Materials Science}}},
+ shorttitle = {Evaluation of the {{MACE Force Field Architecture}}},
+ author = {Kovacs, David Peter and Batatia, Ilyes and Arany, Eszter Sara and Csanyi, Gabor},
+ date = {2023-07-28},
+ eprint = {2305.14247},
+ eprinttype = {arXiv},
+ eprintclass = {physics, stat},
+ doi = {10.1063/5.0155322},
+ url = {http://arxiv.org/abs/2305.14247v2},
+ urldate = {2024-06-17},
+ abstract = {The MACE architecture represents the state of the art in the field of machine learning force fields for a variety of in-domain, extrapolation and low-data regime tasks. In this paper, we further evaluate MACE by fitting models for published benchmark datasets. We show that MACE generally outperforms alternatives for a wide range of systems from amorphous carbon, universal materials modelling, and general small molecule organic chemistry to large molecules and liquid water. We demonstrate the capabilities of the model on tasks ranging from constrained geometry optimisation to molecular dynamics simulations and find excellent performance across all tested domains. We show that MACE is very data efficient, and can reproduce experimental molecular vibrational spectra when trained on as few as 50 randomly selected reference configurations. We further demonstrate that the strictly local atom-centered model is sufficient for such tasks even in the case of large molecules and weakly interacting molecular assemblies.},
+ pubstate = {prepublished},
+ keywords = {/unread,Physics - Chemical Physics,Statistics - Machine Learning},
+ file = {/Users/wasmer/Nextcloud/Zotero/Kovacs et al_2023_Evaluation of the MACE Force Field Architecture2.pdf;/Users/wasmer/Zotero/storage/24BEKIUD/2305.html}
+}
+
@article{kovacsLinearAtomicCluster2021,
title = {Linear {{Atomic Cluster Expansion Force Fields}} for {{Organic Molecules}}: {{Beyond RMSE}}},
shorttitle = {Linear {{Atomic Cluster Expansion Force Fields}} for {{Organic Molecules}}},
@@ -9219,29 +10072,46 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Kovács, Dávid Péter and Moore, J. Harry and Browning, Nicholas J. and Batatia, Ilyes and Horton, Joshua T. and Kapil, Venkat and Magdău, Ioan-Bogdan and Cole, Daniel J. and Csányi, Gábor},
date = {2023-12-23},
eprint = {2312.15211v1},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
url = {https://arxiv.org/abs/2312.15211v1},
urldate = {2024-01-02},
abstract = {Classical empirical force fields have dominated biomolecular simulation for over 50 years. Although widely used in drug discovery, crystal structure prediction, and biomolecular dynamics, they generally lack the accuracy and transferability required for predictive modelling. In this paper, we introduce MACE-OFF23, a transferable force field for organic molecules created using state-of-the-art machine learning technology and first-principles reference data computed with a high level of quantum mechanical theory. MACE-OFF23 demonstrates the remarkable capabilities of local, short-range models by accurately predicting a wide variety of gas and condensed phase properties of molecular systems. It produces accurate, easy-to-converge dihedral torsion scans of unseen molecules, as well as reliable descriptions of molecular crystals and liquids, including quantum nuclear effects. We further demonstrate the capabilities of MACE-OFF23 by determining free energy surfaces in explicit solvent, as well as the folding dynamics of peptides. Finally, we simulate a fully solvated small protein, observing accurate secondary structure and vibrational spectrum. These developments enable first-principles simulations of molecular systems for the broader chemistry community at high accuracy and low computational cost.},
langid = {english},
- pubstate = {preprint},
- keywords = {AML,GNN,ML,ML-FF,MLP,molecules,MPNN,organic chemistry,smal organic molecules,transfer learning},
+ pubstate = {prepublished},
+ keywords = {AML,GNN,MACE,ML,ML-FF,MLP,molecules,MPNN,organic chemistry,smal organic molecules,transfer learning,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Kovács et al_2023_MACE-OFF23.pdf}
}
+@article{kovacsMagnetostaticsMicromagneticsPhysics2022,
+ title = {Magnetostatics and Micromagnetics with Physics Informed Neural Networks},
+ author = {Kovacs, Alexander and Exl, Lukas and Kornell, Alexander and Fischbacher, Johann and Hovorka, Markus and Gusenbauer, Markus and Breth, Leoni and Oezelt, Harald and Praetorius, Dirk and Suess, Dieter and Schrefl, Thomas},
+ date = {2022-04-15},
+ journaltitle = {Journal of Magnetism and Magnetic Materials},
+ shortjournal = {Journal of Magnetism and Magnetic Materials},
+ volume = {548},
+ pages = {168951},
+ issn = {0304-8853},
+ doi = {10.1016/j.jmmm.2021.168951},
+ url = {https://www.sciencedirect.com/science/article/pii/S0304885321011483},
+ urldate = {2024-06-11},
+ abstract = {Partial differential equations and variational problems can be solved with physics informed neural networks (PINNs). The unknown field is approximated with neural networks. Minimizing the residuals of the static Maxwell equation at collocation points or the magnetostatic energy, the weights of the neural network are adjusted so that the neural network solution approximates the magnetic vector potential. This way, the magnetic flux density for a given magnetization distribution can be estimated. With the magnetization as an additional unknown, inverse magnetostatic problems can be solved. Augmenting the magnetostatic energy with additional energy terms, micromagnetic problems can be solved. We demonstrate the use of physics informed neural networks for solving magnetostatic problems, computing the magnetization for inverse problems, and calculating the demagnetization curves for two-dimensional geometries.},
+ keywords = {/unread,AML,continuum physics,inverse problem,magnetism,magnetostatics,micromagnetics,ML,ML-PDE,physics-informed ML,PINN,prediction of magnetic flux,prediction of magnetization,unsupervised learning},
+ file = {/Users/wasmer/Nextcloud/Zotero/Kovacs et al_2022_Magnetostatics and micromagnetics with physics informed neural networks.pdf;/Users/wasmer/Zotero/storage/4QZQ293Q/S0304885321011483.html}
+}
+
@online{krennPredictingFutureAI2022,
title = {Predicting the {{Future}} of {{AI}} with {{AI}}: {{High-quality}} Link Prediction in an Exponentially Growing Knowledge Network},
shorttitle = {Predicting the {{Future}} of {{AI}} with {{AI}}},
author = {Krenn, Mario and Buffoni, Lorenzo and Coutinho, Bruno and Eppel, Sagi and Foster, Jacob Gates and Gritsevskiy, Andrew and Lee, Harlin and Lu, Yichao and Moutinho, Joao P. and Sanjabi, Nima and Sonthalia, Rishi and Tran, Ngoc Mai and Valente, Francisco and Xie, Yangxinyu and Yu, Rose and Kopp, Michael},
date = {2022-09-23},
eprint = {2210.00881},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2210.00881},
url = {http://arxiv.org/abs/2210.00881},
urldate = {2022-10-05},
abstract = {A tool that could suggest new personalized research directions and ideas by taking insights from the scientific literature could significantly accelerate the progress of science. A field that might benefit from such an approach is artificial intelligence (AI) research, where the number of scientific publications has been growing exponentially over the last years, making it challenging for human researchers to keep track of the progress. Here, we use AI techniques to predict the future research directions of AI itself. We develop a new graph-based benchmark based on real-world data -- the Science4Cast benchmark, which aims to predict the future state of an evolving semantic network of AI. For that, we use more than 100,000 research papers and build up a knowledge network with more than 64,000 concept nodes. We then present ten diverse methods to tackle this task, ranging from pure statistical to pure learning methods. Surprisingly, the most powerful methods use a carefully curated set of network features, rather than an end-to-end AI approach. It indicates a great potential that can be unleashed for purely ML approaches without human knowledge. Ultimately, better predictions of new future research directions will be a crucial component of more advanced research suggestion tools.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {General ML,literature analysis},
file = {/Users/wasmer/Nextcloud/Zotero/Krenn et al_2022_Predicting the Future of AI with AI.pdf;/Users/wasmer/Zotero/storage/MZBX2N4K/2210.html}
}
@@ -9334,6 +10204,24 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
file = {/Users/wasmer/Nextcloud/Zotero/Kulik et al_2022_Roadmap on Machine Learning in Electronic Structure4.pdf}
}
+@article{kumagaiInsightsOxygenVacancies2021,
+ title = {Insights into Oxygen Vacancies from High-Throughput First-Principles Calculations},
+ author = {Kumagai, Yu and Tsunoda, Naoki and Takahashi, Akira and Oba, Fumiyasu},
+ date = {2021-12-27},
+ journaltitle = {Physical Review Materials},
+ shortjournal = {Phys. Rev. Mater.},
+ volume = {5},
+ number = {12},
+ pages = {123803},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevMaterials.5.123803},
+ url = {https://link.aps.org/doi/10.1103/PhysRevMaterials.5.123803},
+ urldate = {2024-05-24},
+ abstract = {Oxygen vacancies play significant roles in various properties of oxide materials. Therefore, insights into the oxygen vacancies can facilitate the discovery of better oxide materials. To achieve this, we developed codes for high-throughput point-defect calculations and applied them to characterize oxygen vacancies in 937 oxides. From the resulting large dataset, we analyzed the vacancy structures and formation energies and constructed machine-learning regression models to predict vacancy formation energies. We have found that the vacancy formation energies are predicted using the random forest regression models with accuracies of 0.27–0.44 eV depending on the charge states. Analyses of the importance of the descriptors show that the formation energies of the neutral vacancies are mainly determined by the orbital characteristics of the conduction-band minima, the oxide stability, and the band gaps, whereas those of the doubly charged defects are determined by factors related to electrostatic energy. These codes and datasets are publicly available, and a graphical user interface is available to analyze the calculation results.},
+ keywords = {/unread,AML,charged defects,compositional descriptors,defects,descriptors,DFT,Fireworks,library,materials,ML,oxides,physics,point defects,Python,random forest,scikit-learn,vacancies,with-code,workflows},
+ file = {/Users/wasmer/Nextcloud/Zotero/Kumagai et al_2021_Insights into oxygen vacancies from high-throughput first-principles.pdf;/Users/wasmer/Zotero/storage/P6IJISMH/PhysRevMaterials.5.html}
+}
+
@article{kumarTopologicalQuantumMaterials2021,
title = {Topological {{Quantum Materials}} from the {{Viewpoint}} of {{Chemistry}}},
author = {Kumar, Nitesh and Guin, Satya N. and Manna, Kaustuv and Shekhar, Chandra and Felser, Claudia},
@@ -9349,7 +10237,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
url = {https://doi.org/10.1021/acs.chemrev.0c00732},
urldate = {2021-05-13},
abstract = {Topology, a mathematical concept, has recently become a popular and truly transdisciplinary topic encompassing condensed matter physics, solid state chemistry, and materials science. Since there is a direct connection between real space, namely atoms, valence electrons, bonds, and orbitals, and reciprocal space, namely bands and Fermi surfaces, via symmetry and topology, classifying topological materials within a single-particle picture is possible. Currently, most materials are classified as trivial insulators, semimetals, and metals or as topological insulators, Dirac and Weyl nodal-line semimetals, and topological metals. The key ingredients for topology are certain symmetries, the inert pair effect of the outer electrons leading to inversion of the conduction and valence bands, and spin–orbit coupling. This review presents the topological concepts related to solids from the viewpoint of a solid-state chemist, summarizes techniques for growing single crystals, and describes basic physical property measurement techniques to characterize topological materials beyond their structure and provide examples of such materials. Finally, a brief outlook on the impact of topology in other areas of chemistry is provided at the end of the article.},
- keywords = {\_tablet,chemistry,topological insulator},
+ keywords = {chemistry,condensed matter,good figures,physics,quantum materials,review,review-of-TIs,topological,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Kumar et al_2021_Topological Quantum Materials from the Viewpoint of Chemistry.pdf}
}
@@ -9372,20 +10260,38 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
file = {/Users/wasmer/Nextcloud/Zotero/Kurth_Perdew_2000_Role of the exchange–correlation energy.pdf;/Users/wasmer/Zotero/storage/EVE2LZLC/(SICI)1097-461X(2000)775814AID-QUA33.0.html}
}
+@article{kurtzRevisitingTerawattChallenge2020,
+ title = {Revisiting the {{Terawatt Challenge}}},
+ author = {Kurtz, Sarah R. and Leilaeioun, Ashling Mehdi and King, Richard R. and Peters, Ian Marius and Heben, Michael J. and Metzger, Wyatt K. and Haegel, Nancy M.},
+ date = {2020-03-01},
+ journaltitle = {MRS Bulletin},
+ shortjournal = {MRS Bulletin},
+ volume = {45},
+ number = {3},
+ pages = {159--164},
+ issn = {1938-1425},
+ doi = {10.1557/mrs.2020.73},
+ url = {https://doi.org/10.1557/mrs.2020.73},
+ urldate = {2024-08-01},
+ langid = {english},
+ keywords = {/unread,energy challenge,energy materials,for introductions,solar energy},
+ file = {/Users/wasmer/Nextcloud/Zotero/Kurtz et al_2020_Revisiting the Terawatt Challenge.pdf}
+}
+
@online{laaksoUpdatesDScribeLibrary2023,
title = {Updates to the {{DScribe Library}}: {{New Descriptors}} and {{Derivatives}}},
shorttitle = {Updates to the {{DScribe Library}}},
author = {Laakso, Jarno and Himanen, Lauri and Homm, Henrietta and Morooka, Eiaki V. and Jäger, Marc O. J. and Todorović, Milica and Rinke, Patrick},
date = {2023-03-24},
eprint = {2303.14046},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2303.14046},
url = {http://arxiv.org/abs/2303.14046},
urldate = {2023-06-12},
abstract = {We present an update of the DScribe package, a Python library for atomistic descriptors. The update extends DScribe's descriptor selection with the Valle-Oganov materials fingerprint and provides descriptor derivatives to enable more advanced machine learning tasks, such as force prediction and structure optimization. For all descriptors, numeric derivatives are now available in DSribe. For the many-body tensor representation (MBTR) and the Smooth Overlap of Atomic Positions (SOAP), we have also implemented analytic derivatives. We demonstrate the effectiveness of the descriptor derivatives for machine learning models of Cu clusters and perovskite alloys.},
- pubstate = {preprint},
- keywords = {/unread,\_tablet,ACSF,analytical derivatives,derivatives,descriptors,DScribe,library,materials,Matrix descriptors,MBTR,ML,Open source,Python,rec-by-ruess,SOAP,Valle-Oganov descriptor,with-code},
+ pubstate = {prepublished},
+ keywords = {/unread,ACSF,analytical derivatives,derivatives,descriptors,DScribe,library,materials,Matrix descriptors,MBTR,ML,Open source,Python,rec-by-ruess,SOAP,Valle-Oganov descriptor,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Laakso et al_2023_Updates to the DScribe Library.pdf;/Users/wasmer/Zotero/storage/JNKQJJYE/2303.html}
}
@@ -9394,30 +10300,46 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Labrie-Boulay, Isaac and Winkler, Thomas Brian and Franzen, Daniel and Romanova, Alena and Fangohr, Hans and Kläui, Mathias},
date = {2023-03-24},
eprint = {2303.16905},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2303.16905},
url = {http://arxiv.org/abs/2303.16905},
urldate = {2023-04-12},
abstract = {One of the most important magnetic spin structure is the topologically stabilised skyrmion quasi-particle. Its interesting physical properties make them candidates for memory and efficient neuromorphic computation schemes. For the device operation, detection of the position, shape, and size of skyrmions is required and magnetic imaging is typically employed. A frequently used technique is magneto-optical Kerr microscopy where depending on the samples material composition, temperature, material growing procedures, etc., the measurements suffer from noise, low-contrast, intensity gradients, or other optical artifacts. Conventional image analysis packages require manual treatment, and a more automatic solution is required. We report a convolutional neural network specifically designed for segmentation problems to detect the position and shape of skyrmions in our measurements. The network is tuned using selected techniques to optimize predictions and in particular the number of detected classes is found to govern the performance. The results of this study shows that a well-trained network is a viable method of automating data pre-processing in magnetic microscopy. The approach is easily extendable to other spin structures and other magnetic imaging methods.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {CNN,computer vision,Dzyaloshinskii–Moriya interaction,experimental science,image segmentation,magnetism,ML,object detection,skyrmions,Spintronics},
file = {/Users/wasmer/Nextcloud/Zotero/Labrie-Boulay et al_2023_Machine learning-based spin structure detection.pdf;/Users/wasmer/Zotero/storage/SEU5QPUQ/2303.html}
}
+@article{lafuente-bartolomeTopologicalPolaronsHalide2024,
+ title = {Topological Polarons in Halide Perovskites},
+ author = {Lafuente-Bartolome, Jon and Lian, Chao and Giustino, Feliciano},
+ date = {2024-05-21},
+ journaltitle = {Proceedings of the National Academy of Sciences},
+ volume = {121},
+ number = {21},
+ pages = {e2318151121},
+ publisher = {Proceedings of the National Academy of Sciences},
+ doi = {10.1073/pnas.2318151121},
+ url = {https://www.pnas.org/doi/10.1073/pnas.2318151121},
+ urldate = {2024-07-05},
+ abstract = {Halide perovskites emerged as a revolutionary family of high-quality semiconductors for solar energy harvesting and energy-efficient lighting. There is mounting evidence that the exceptional optoelectronic properties of these materials could stem from unconventional electron–phonon couplings, and it has been suggested that the formation of polarons and self-trapped excitons could be key to understanding such properties. By performing first-principles simulations across the length scales, here we show that halide perovskites harbor a uniquely rich variety of polaronic species, including small polarons, large polarons, and charge density waves, and we explain a variety of experimental observations. We find that these emergent quasiparticles support topologically nontrivial phonon fields with quantized topological charge, making them nonmagnetic analog of the helical Bloch points found in magnetic skyrmion lattices.},
+ keywords = {/unread,DFT,energy efficiency,energy materials,halovskites,large simulations,physics,polaron,Quantum ESPRESSO,skyrmions,Supercomputer,topological}
+}
+
@online{lakshminarayananSimpleScalablePredictive2017,
title = {Simple and {{Scalable Predictive Uncertainty Estimation}} Using {{Deep Ensembles}}},
author = {Lakshminarayanan, Balaji and Pritzel, Alexander and Blundell, Charles},
date = {2017-11-03},
eprint = {1612.01474},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.1612.01474},
url = {http://arxiv.org/abs/1612.01474},
urldate = {2023-12-05},
abstract = {Deep neural networks (NNs) are powerful black box predictors that have recently achieved impressive performance on a wide spectrum of tasks. Quantifying predictive uncertainty in NNs is a challenging and yet unsolved problem. Bayesian NNs, which learn a distribution over weights, are currently the state-of-the-art for estimating predictive uncertainty; however these require significant modifications to the training procedure and are computationally expensive compared to standard (non-Bayesian) NNs. We propose an alternative to Bayesian NNs that is simple to implement, readily parallelizable, requires very little hyperparameter tuning, and yields high quality predictive uncertainty estimates. Through a series of experiments on classification and regression benchmarks, we demonstrate that our method produces well-calibrated uncertainty estimates which are as good or better than approximate Bayesian NNs. To assess robustness to dataset shift, we evaluate the predictive uncertainty on test examples from known and unknown distributions, and show that our method is able to express higher uncertainty on out-of-distribution examples. We demonstrate the scalability of our method by evaluating predictive uncertainty estimates on ImageNet.},
- pubstate = {preprint},
- keywords = {/unread,active learning,deep ensembles,ensemble learning,General ML,library,ML,original publication,uncertainty quantification,with-code},
+ pubstate = {prepublished},
+ keywords = {active learning,deep ensembles,DeepMind,ensemble learning,General ML,library,ML,original publication,uncertainty quantification,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Lakshminarayanan et al_2017_Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles.pdf;/Users/wasmer/Zotero/storage/AV8HUIKB/1612.html}
}
@@ -9427,29 +10349,46 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lam, Remi and Sanchez-Gonzalez, Alvaro and Willson, Matthew and Wirnsberger, Peter and Fortunato, Meire and Pritzel, Alexander and Ravuri, Suman and Ewalds, Timo and Alet, Ferran and Eaton-Rosen, Zach and Hu, Weihua and Merose, Alexander and Hoyer, Stephan and Holland, George and Stott, Jacklynn and Vinyals, Oriol and Mohamed, Shakir and Battaglia, Peter},
date = {2022-12-24},
eprint = {2212.12794},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2212.12794},
url = {http://arxiv.org/abs/2212.12794},
urldate = {2022-12-31},
abstract = {We introduce a machine-learning (ML)-based weather simulator--called "GraphCast"--which outperforms the most accurate deterministic operational medium-range weather forecasting system in the world, as well as all previous ML baselines. GraphCast is an autoregressive model, based on graph neural networks and a novel high-resolution multi-scale mesh representation, which we trained on historical weather data from the European Centre for Medium-Range Weather Forecasts (ECMWF)'s ERA5 reanalysis archive. It can make 10-day forecasts, at 6-hour time intervals, of five surface variables and six atmospheric variables, each at 37 vertical pressure levels, on a 0.25-degree latitude-longitude grid, which corresponds to roughly 25 x 25 kilometer resolution at the equator. Our results show GraphCast is more accurate than ECMWF's deterministic operational forecasting system, HRES, on 90.0\% of the 2760 variable and lead time combinations we evaluated. GraphCast also outperforms the most accurate previous ML-based weather forecasting model on 99.2\% of the 252 targets it reported. GraphCast can generate a 10-day forecast (35 gigabytes of data) in under 60 seconds on Cloud TPU v4 hardware. Unlike traditional forecasting methods, ML-based forecasting scales well with data: by training on bigger, higher quality, and more recent data, the skill of the forecasts can improve. Together these results represent a key step forward in complementing and improving weather modeling with ML, open new opportunities for fast, accurate forecasting, and help realize the promise of ML-based simulation in the physical sciences.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,DeepMind,GNN,PDE,SciML,simulation,weather forecasting},
file = {/Users/wasmer/Nextcloud/Zotero/Lam et al_2022_GraphCast.pdf;/Users/wasmer/Zotero/storage/8UD54ESE/2212.html}
}
+@online{lamGraphCastLearningSkillful2023,
+ title = {{{GraphCast}}: {{Learning}} Skillful Medium-Range Global Weather Forecasting},
+ shorttitle = {{{GraphCast}}},
+ author = {Lam, Remi and Sanchez-Gonzalez, Alvaro and Willson, Matthew and Wirnsberger, Peter and Fortunato, Meire and Alet, Ferran and Ravuri, Suman and Ewalds, Timo and Eaton-Rosen, Zach and Hu, Weihua and Merose, Alexander and Hoyer, Stephan and Holland, George and Vinyals, Oriol and Stott, Jacklynn and Pritzel, Alexander and Mohamed, Shakir and Battaglia, Peter},
+ date = {2023-08-04},
+ eprint = {2212.12794},
+ eprinttype = {arXiv},
+ eprintclass = {physics},
+ doi = {10.48550/arXiv.2212.12794},
+ url = {http://arxiv.org/abs/2212.12794},
+ urldate = {2024-08-02},
+ abstract = {Global medium-range weather forecasting is critical to decision-making across many social and economic domains. Traditional numerical weather prediction uses increased compute resources to improve forecast accuracy, but cannot directly use historical weather data to improve the underlying model. We introduce a machine learning-based method called "GraphCast", which can be trained directly from reanalysis data. It predicts hundreds of weather variables, over 10 days at 0.25 degree resolution globally, in under one minute. We show that GraphCast significantly outperforms the most accurate operational deterministic systems on 90\% of 1380 verification targets, and its forecasts support better severe event prediction, including tropical cyclones, atmospheric rivers, and extreme temperatures. GraphCast is a key advance in accurate and efficient weather forecasting, and helps realize the promise of machine learning for modeling complex dynamical systems.},
+ pubstate = {prepublished},
+ keywords = {/unread,AI4Science,DeepMind,GNN,weather forecasting,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Lam et al_2023_GraphCast.pdf;/Users/wasmer/Zotero/storage/W4M6UQC8/2212.html}
+}
+
@online{langeNeuralNetworkApproach2023,
title = {Neural Network Approach to Quasiparticle Dispersions in Doped Antiferromagnets},
author = {Lange, Hannah and Döschl, Fabian and Carrasquilla, Juan and Bohrdt, Annabelle},
date = {2023-10-12},
eprint = {2310.08578},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:quant-ph},
doi = {10.48550/arXiv.2310.08578},
url = {http://arxiv.org/abs/2310.08578},
urldate = {2024-02-28},
abstract = {Numerically simulating spinful, fermionic systems is of great interest from the perspective of condensed matter physics. However, the exponential growth of the Hilbert space dimension with system size renders an exact parameterization of large quantum systems prohibitively demanding. This is a perfect playground for neural networks, owing to their immense representative power that often allows to use only a fraction of the parameters that are needed in exact methods. Here, we investigate the ability of neural quantum states (NQS) to represent the bosonic and fermionic \$t-J\$ model - the high interaction limit of the Fermi-Hubbard model - on different 1D and 2D lattices. Using autoregressive recurrent neural networks (RNNs) with 2D tensorized gated recurrent units, we study the ground state representations upon doping the half-filled system with holes. Moreover, we present a method to calculate dispersion relations from the neural network state representation, applicable to any neural network architecture and any lattice geometry, that allows to infer the low-energy physics from the NQS. To demonstrate our approach, we calculate the dispersion of a single hole in the \$t-J\$ model on different 1D and 2D square and triangular lattices. Furthermore, we analyze the strengths and weaknesses of the RNN approach for fermionic systems, pointing the way for an accurate and efficient parameterization of fermionic quantum systems using neural quantum states.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,Heisenberg model,ML,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Lange et al_2023_Neural network approach to quasiparticle dispersions in doped antiferromagnets.pdf;/Users/wasmer/Zotero/storage/8FDGYN7V/2310.html}
}
@@ -9459,28 +10398,44 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Langer, Marcel F. and Knoop, Florian and Carbogno, Christian and Scheffler, Matthias and Rupp, Matthias},
date = {2023-03-28},
eprint = {2303.14434},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2303.14434},
url = {http://arxiv.org/abs/2303.14434},
urldate = {2023-04-04},
abstract = {The Green-Kubo (GK) method is a rigorous framework for heat transport simulations in materials. However, it requires an accurate description of the potential-energy surface and carefully converged statistics. Machine-learning potentials can achieve the accuracy of first-principles simulations while allowing to reach well beyond their simulation time and length scales at a fraction of the cost. In this paper, we explain how to apply the GK approach to the recent class of message-passing machine-learning potentials, which iteratively consider semi-local interactions beyond the initial interaction cutoff. We derive an adapted heat flux formulation that can be implemented using automatic differentiation without compromising computational efficiency. The approach is demonstrated and validated by calculating the thermal conductivity of zirconium dioxide across temperatures.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,autodiff,GKNet,GNN,Gree-Kubo,JAX,long-range interaction,materials,MD,ML,MLP,MPNN,SchNet,SchNetPack,semilocal interactions,tensorial target,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Langer et al_2023_Heat flux for semi-local machine-learning potentials.pdf;/Users/wasmer/Zotero/storage/Q6SAW6JE/2303.html}
}
+@online{langerProbingEffectsBroken2024,
+ title = {Probing the Effects of Broken Symmetries in Machine Learning},
+ author = {Langer, Marcel F. and Pozdnyakov, Sergey N. and Ceriotti, Michele},
+ date = {2024-06-25},
+ eprint = {2406.17747},
+ eprinttype = {arXiv},
+ eprintclass = {physics, stat},
+ doi = {10.48550/arXiv.2406.17747},
+ url = {http://arxiv.org/abs/2406.17747},
+ urldate = {2024-06-27},
+ abstract = {Symmetry is one of the most central concepts in physics, and it is no surprise that it has also been widely adopted as an inductive bias for machine-learning models applied to the physical sciences. This is especially true for models targeting the properties of matter at the atomic scale. Both established and state-of-the-art approaches, with almost no exceptions, are built to be exactly equivariant to translations, permutations, and rotations of the atoms. Incorporating symmetries -- rotations in particular -- constrains the model design space and implies more complicated architectures that are often also computationally demanding. There are indications that non-symmetric models can easily learn symmetries from data, and that doing so can even be beneficial for the accuracy of the model. We put a model that obeys rotational invariance only approximately to the test, in realistic scenarios involving simulations of gas-phase, liquid, and solid water. We focus specifically on physical observables that are likely to be affected -- directly or indirectly -- by symmetry breaking, finding negligible consequences when the model is used in an interpolative, bulk, regime. Even for extrapolative gas-phase predictions, the model remains very stable, even though symmetry artifacts are noticeable. We also discuss strategies that can be used to systematically reduce the magnitude of symmetry breaking when it occurs, and assess their impact on the convergence of observables.},
+ pubstate = {prepublished},
+ keywords = {/unread,AML,approximative equivariance,equivariant alternative,ML,MLP,transformer},
+ file = {/Users/wasmer/Nextcloud/Zotero/Langer et al_2024_Probing the effects of broken symmetries in machine learning.pdf;/Users/wasmer/Zotero/storage/WDB6G79D/2406.html}
+}
+
@unpublished{langerRepresentationsMoleculesMaterials2021,
title = {Representations of Molecules and Materials for Interpolation of Quantum-Mechanical Simulations via Machine Learning},
author = {Langer, Marcel F. and Goeßmann, Alex and Rupp, Matthias},
date = {2021-02-09},
eprint = {2003.12081},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2003.12081},
urldate = {2021-05-13},
abstract = {Computational study of molecules and materials from first principles is a cornerstone of physics, chemistry, and materials science, but limited by the cost of accurate and precise simulations. In settings involving many simulations, machine learning can reduce these costs, often by orders of magnitude, by interpolating between reference simulations. This requires representations that describe any molecule or material and support interpolation. We comprehensively review and discuss current representations and relations between them, using a unified mathematical framework based on many-body functions, group averaging, and tensor products. For selected state-of-the-art representations, we compare energy predictions for organic molecules, binary alloys, and Al-Ga-In sesquioxides in numerical experiments controlled for data distribution, regression method, and hyper-parameter optimization.},
- keywords = {\_tablet,ACE,benchmarking,BoB,BS,CM,descriptors,GPR,KRR,library,materials,MBTR,ML,models,MTP,review,SOAP,with-code},
+ keywords = {ACE,benchmarking,BoB,BS,CM,descriptors,GPR,KRR,library,materials,MBTR,ML,models,MTP,review,SOAP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Langer et al_2021_Representations of molecules and materials for interpolation of.pdf;/Users/wasmer/Zotero/storage/5BG77UWY/2003.html}
}
@@ -9501,7 +10456,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
abstract = {Computational study of molecules and materials from first principles is a cornerstone of physics, chemistry, and materials science, but limited by the cost of accurate and precise simulations. In settings involving many simulations, machine learning can reduce these costs, often by orders of magnitude, by interpolating between reference simulations. This requires representations that describe any molecule or material and support interpolation. We comprehensively review and discuss current representations and relations between them. For selected state-of-the-art representations, we compare energy predictions for organic molecules, binary alloys, and Al–Ga–In sesquioxides in numerical experiments controlled for data distribution, regression method, and hyper-parameter optimization.},
issue = {1},
langid = {english},
- keywords = {\_tablet,ACE,autoML,benchmarking,BoB,BS,CM,descriptor comparison,descriptors,GPR,hyperparameters optimization,KRR,library,materials,MBTR,ML,model reporting,models,MTP,review,SOAP,with-code},
+ keywords = {ACE,autoML,benchmarking,BoB,BS,CM,descriptor comparison,descriptors,GPR,hyperparameters optimization,KRR,library,materials,MBTR,ML,model reporting,models,MTP,review,SOAP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Langer et al_2022_Representations of molecules and materials for interpolation of.pdf;/Users/wasmer/Nextcloud/Zotero/Langer et al_2022_Representations of molecules and materials for interpolation of2.pdf;/Users/wasmer/Zotero/storage/9RVUDSSX/s41524-022-00721-x.html}
}
@@ -9548,13 +10503,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lászlóffy, András and Nyári, Bendegúz and Csire, Gábor and Szunyogh, László and Újfalussy, Balázs},
date = {2023-08-26},
eprint = {2308.13831},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2308.13831},
url = {http://arxiv.org/abs/2308.13831},
urldate = {2023-09-20},
abstract = {Recent advances in electron spin resonance techniques have allowed the manipulation of the spin of individual atoms, making magnetic atomic chains on superconducting hosts one of the most promising platform where topological superconductivity can be engineered. Motivated by this progress, we provide a detailed, quantitative description of the effects of manipulating spins in realistic nanowires by applying a first-principles-based computational approach to a recent experiment: an iron chain deposited on top of Au/Nb heterostructure. As a continuation of the first part of the paper experimentally relevant computational experiments are performed in spin spiral chains that shed light on several concerns about practical applications and add new aspects to the interpretation of recent experiments. We explore the stability of topological zero energy states, the formation and distinction of topologically trivial and non-trivial zero energy edge states, the effect of local changes in the exchange fields, the emergence of topological fragmentation, and the shift of Majorana Zero Modes along the superconducting nanowires opening avenues toward the implementation of a braiding operation.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,Budapest KKR group,GF2023 workshop,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Lászlóffy et al_2023_Topological superconductivity from first-principles II.pdf;/Users/wasmer/Zotero/storage/WPNB3PLJ/2308.html}
}
@@ -9565,13 +10520,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lavin, Alexander and Krakauer, David and Zenil, Hector and Gottschlich, Justin and Mattson, Tim and Brehmer, Johann and Anandkumar, Anima and Choudry, Sanjay and Rocki, Kamil and Baydin, Atılım Güneş and Prunkl, Carina and Paige, Brooks and Isayev, Olexandr and Peterson, Erik and McMahon, Peter L. and Macke, Jakob and Cranmer, Kyle and Zhang, Jiaxin and Wainwright, Haruko and Hanuka, Adi and Veloso, Manuela and Assefa, Samuel and Zheng, Stephan and Pfeffer, Avi},
date = {2022-11-27},
eprint = {2112.03235},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2112.03235},
url = {http://arxiv.org/abs/2112.03235},
urldate = {2023-08-21},
abstract = {The original "Seven Motifs" set forth a roadmap of essential methods for the field of scientific computing, where a motif is an algorithmic method that captures a pattern of computation and data movement. We present the "Nine Motifs of Simulation Intelligence", a roadmap for the development and integration of the essential algorithms necessary for a merger of scientific computing, scientific simulation, and artificial intelligence. We call this merger simulation intelligence (SI), for short. We argue the motifs of simulation intelligence are interconnected and interdependent, much like the components within the layers of an operating system. Using this metaphor, we explore the nature of each layer of the simulation intelligence operating system stack (SI-stack) and the motifs therein: (1) Multi-physics and multi-scale modeling; (2) Surrogate modeling and emulation; (3) Simulation-based inference; (4) Causal modeling and inference; (5) Agent-based modeling; (6) Probabilistic programming; (7) Differentiable programming; (8) Open-ended optimization; (9) Machine programming. We believe coordinated efforts between motifs offers immense opportunity to accelerate scientific discovery, from solving inverse problems in synthetic biology and climate science, to directing nuclear energy experiments and predicting emergent behavior in socioeconomic settings. We elaborate on each layer of the SI-stack, detailing the state-of-art methods, presenting examples to highlight challenges and opportunities, and advocating for specific ways to advance the motifs and the synergies from their combinations. Advancing and integrating these technologies can enable a robust and efficient hypothesis-simulation-analysis type of scientific method, which we introduce with several use-cases for human-machine teaming and automated science.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Pasteur \& ISI,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Lavin et al_2022_Simulation Intelligence.pdf;/Users/wasmer/Zotero/storage/LYDGASRK/2112.html}
}
@@ -9628,13 +10583,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lee, Namkyeong and Noh, Heewoong and Kim, Sungwon and Hyun, Dongmin and Na, Gyoung S. and Park, Chanyoung},
date = {2023-04-10},
eprint = {2303.07000},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2303.07000},
url = {http://arxiv.org/abs/2303.07000},
urldate = {2023-09-23},
abstract = {The density of states (DOS) is a spectral property of materials, which provides fundamental insights on various characteristics of materials. In this paper, we propose a model to predict the DOS by reflecting the nature of DOS: DOS determines the general distribution of states as a function of energy. Specifically, we integrate the heterogeneous information obtained from the crystal structure and the energies via multi-modal transformer, thereby modeling the complex relationships between the atoms in the crystal structure, and various energy levels. Extensive experiments on two types of DOS, i.e., Phonon DOS and Electron DOS, with various real-world scenarios demonstrate the superiority of DOSTransformer. The source code for DOSTransformer is available at https://github.com/HeewoongNoh/DOSTransformer.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,Computer Science - Machine Learning,Condensed Matter - Materials Science,Physics - Computational Physics},
file = {/Users/wasmer/Zotero/storage/DG95U3QT/Lee et al. - 2023 - Predicting Density of States via Multi-modal Trans.pdf;/Users/wasmer/Zotero/storage/JTRZEV35/2303.html}
}
@@ -9796,13 +10751,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lessig, Christian and Luise, Ilaria and Gong, Bing and Langguth, Michael and Stadtler, Scarlet and Schultz, Martin},
date = {2023-09-07},
eprint = {2308.13280},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2308.13280},
url = {http://arxiv.org/abs/2308.13280},
urldate = {2023-11-12},
abstract = {The atmosphere affects humans in a multitude of ways, from loss of life due to adverse weather effects to long-term social and economic impacts on societies. Computer simulations of atmospheric dynamics are, therefore, of great importance for the well-being of our and future generations. Here, we propose AtmoRep, a novel, task-independent stochastic computer model of atmospheric dynamics that can provide skillful results for a wide range of applications. AtmoRep uses large-scale representation learning from artificial intelligence to determine a general description of the highly complex, stochastic dynamics of the atmosphere from the best available estimate of the system's historical trajectory as constrained by observations. This is enabled by a novel self-supervised learning objective and a unique ensemble that samples from the stochastic model with a variability informed by the one in the historical record. The task-independent nature of AtmoRep enables skillful results for a diverse set of applications without specifically training for them and we demonstrate this for nowcasting, temporal interpolation, model correction, and counterfactuals. We also show that AtmoRep can be improved with additional data, for example radar observations, and that it can be extended to tasks such as downscaling. Our work establishes that large-scale neural networks can provide skillful, task-independent models of atmospheric dynamics. With this, they provide a novel means to make the large record of atmospheric observations accessible for applications and for scientific inquiry, complementing existing simulations based on first principles.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AI4Science,atmospheric physics,dynamical systems,foundation models,FZJ,generative models,JSC,masked token model,multiformer,multimodal input,pretrained models,representation learning,self-attention,SSL,stochastic modeling,transformer,weather forecasting},
file = {/Users/wasmer/Nextcloud/Zotero/Lessig et al_2023_AtmoRep.pdf;/Users/wasmer/Zotero/storage/27DMQXFS/2308.html}
}
@@ -9812,7 +10767,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lewis, Alan M. and Grisafi, Andrea and Ceriotti, Michele and Rossi, Mariana},
date = {2021-06-09},
eprint = {2106.05364},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2106.05364},
urldate = {2021-06-29},
@@ -9836,22 +10791,40 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
url = {https://doi.org/10.1021/acs.jctc.1c00576},
urldate = {2022-08-22},
abstract = {We introduce a local machine-learning method for predicting the electron densities of periodic systems. The framework is based on a numerical, atom-centered auxiliary basis, which enables an accurate expansion of the all-electron density in a form suitable for learning isolated and periodic systems alike. We show that, using this formulation, the electron densities of metals, semiconductors, and molecular crystals can all be accurately predicted using symmetry-adapted Gaussian process regression models, properly adjusted for the nonorthogonal nature of the basis. These predicted densities enable the efficient calculation of electronic properties, which present errors on the order of tens of meV/atom when compared to ab initio density-functional calculations. We demonstrate the key power of this approach by using a model trained on ice unit cells containing only 4 water molecules to predict the electron densities of cells containing up to 512 molecules and see no increase in the magnitude of the errors of derived electronic properties when increasing the system size. Indeed, we find that these extrapolated derived energies are more accurate than those predicted using a direct machine-learning model. Finally, on heterogeneous data sets SALTED can predict electron densities with errors below 4\%.},
- keywords = {\_tablet,DFT,GPR,lambda-SOAP,library,ML,ML-DFT,ML-ESM,models,molecules,molecules \& solids,prediction of electron density,prediction of ground-state properties,Resolution of the identity,SA-GPR,SALTED,solids,with-code},
+ keywords = {DFT,GPR,lambda-SOAP,library,ML,ML-DFT,ML-ESM,models,molecules,molecules \& solids,prediction of electron density,prediction of ground-state properties,Resolution of the identity,SA-GPR,SALTED,solids,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Lewis et al_2021_Learning Electron Densities in the Condensed Phase.pdf;/Users/wasmer/Zotero/storage/S9FT2FEZ/acs.jctc.html}
}
+@article{lewisPoweringPlanetChemical2006,
+ title = {Powering the Planet: {{Chemical}} Challenges in Solar Energy Utilization},
+ shorttitle = {Powering the Planet},
+ author = {Lewis, Nathan S. and Nocera, Daniel G.},
+ date = {2006-10-24},
+ journaltitle = {Proceedings of the National Academy of Sciences},
+ volume = {103},
+ number = {43},
+ pages = {15729--15735},
+ publisher = {Proceedings of the National Academy of Sciences},
+ doi = {10.1073/pnas.0603395103},
+ url = {https://www.pnas.org/doi/full/10.1073/pnas.0603395103},
+ urldate = {2024-08-01},
+ abstract = {Global energy consumption is projected to increase, even in the face of substantial declines in energy intensity, at least 2-fold by midcentury relative to the present because of population and economic growth. This demand could be met, in principle, from fossil energy resources, particularly coal. However, the cumulative nature of CO2 emissions in the atmosphere demands that holding atmospheric CO2 levels to even twice their preanthropogenic values by midcentury will require invention, development, and deployment of schemes for carbon-neutral energy production on a scale commensurate with, or larger than, the entire present-day energy supply from all sources combined. Among renewable energy resources, solar energy is by far the largest exploitable resource, providing more energy in 1 hour to the earth than all of the energy consumed by humans in an entire year. In view of the intermittency of insolation, if solar energy is to be a major primary energy source, it must be stored and dispatched on demand to the end user. An especially attractive approach is to store solar-converted energy in the form of chemical bonds, i.e., in a photosynthetic process at a year-round average efficiency significantly higher than current plants or algae, to reduce land-area requirements. Scientific challenges involved with this process include schemes to capture and convert solar energy and then store the energy in the form of chemical bonds, producing oxygen from water and a reduced fuel such as hydrogen, methane, methanol, or other hydrocarbon species.},
+ keywords = {chemistry,energy challenge,energy materials,for introductions,solar energy},
+ file = {/Users/wasmer/Nextcloud/Zotero/Lewis_Nocera_2006_Powering the planet.pdf}
+}
+
@online{lewisPredictingElectronicDensity2023,
title = {Predicting the {{Electronic Density Response}} of {{Condensed-Phase Systems}} to {{Electric Field Perturbations}}},
author = {Lewis, Alan M. and Lazzaroni, Paolo and Rossi, Mariana},
date = {2023-04-18},
eprint = {2304.09057},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2304.09057},
url = {http://arxiv.org/abs/2304.09057},
urldate = {2023-05-26},
abstract = {We present a local and transferable machine learning approach capable of predicting the real-space density response of both molecules and periodic systems to external homogeneous electric fields. The new method, SALTER, builds on the Symmetry-Adapted Gaussian Process Regression SALTED framework. SALTER requires only a small, but necessary, modification to the descriptors used to represent the atomic environments. We present the performance of the method on isolated water molecules, bulk water and a naphthalene crystal. Root mean square errors of the predicted density response lie at or below 10\% with barely more than 100 training structures. Derived quantities, such as polarizability tensors and even Raman spectra further derived from these tensors show a good agreement with those calculated directly from quantum mechanical methods. Therefore, SALTER shows excellent performance when predicting derived quantities, while retaining all of the information contained in the full electronic response. This method is thus capable of learning vector fields in a chemical context and serves as a landmark for further developments.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,Electric field,equivariant,GPR,lambda-SOAP,materials,ML,ML-DFT,ML-ESM,molecules,prediction of electron density,prediction of electron response,SA-GPR,SALTED,SOAP},
file = {/Users/wasmer/Nextcloud/Zotero/Lewis et al_2023_Predicting the Electronic Density Response of Condensed-Phase Systems to.pdf;/Users/wasmer/Zotero/storage/6CSD7WWL/2304.html}
}
@@ -9862,13 +10835,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Liao, Yi-Lun and Smidt, Tess},
date = {2023-02-27},
eprint = {2206.11990},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2206.11990},
url = {http://arxiv.org/abs/2206.11990},
urldate = {2023-05-26},
abstract = {Despite their widespread success in various domains, Transformer networks have yet to perform well across datasets in the domain of 3D atomistic graphs such as molecules even when 3D-related inductive biases like translational invariance and rotational equivariance are considered. In this paper, we demonstrate that Transformers can generalize well to 3D atomistic graphs and present Equiformer, a graph neural network leveraging the strength of Transformer architectures and incorporating SE(3)/E(3)-equivariant features based on irreducible representations (irreps). First, we propose a simple and effective architecture by only replacing original operations in Transformers with their equivariant counterparts and including tensor products. Using equivariant operations enables encoding equivariant information in channels of irreps features without complicating graph structures. With minimal modifications to Transformers, this architecture has already achieved strong empirical results. Second, we propose a novel attention mechanism called equivariant graph attention, which improves upon typical attention in Transformers through replacing dot product attention with multi-layer perceptron attention and including non-linear message passing. With these two innovations, Equiformer achieves competitive results to previous models on QM9, MD17 and OC20 datasets.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,attention,E(3),e3nn,equivariant,GNN,graph attention,irreps,MD17,ML,MPNN,OC20,PyG,PyTorch,QM9,SE(3),transformer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Liao_Smidt_2023_Equiformer.pdf;/Users/wasmer/Zotero/storage/33AF9I34/2206.html}
}
@@ -9879,13 +10852,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Liao, Yi-Lun and Wood, Brandon and Das, Abhishek and Smidt, Tess},
date = {2023-06-21},
eprint = {2306.12059},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2306.12059},
url = {http://arxiv.org/abs/2306.12059},
urldate = {2023-08-19},
abstract = {Equivariant Transformers such as Equiformer have demonstrated the efficacy of applying Transformers to the domain of 3D atomistic systems. However, they are still limited to small degrees of equivariant representations due to their computational complexity. In this paper, we investigate whether these architectures can scale well to higher degrees. Starting from Equiformer, we first replace \$SO(3)\$ convolutions with eSCN convolutions to efficiently incorporate higher-degree tensors. Then, to better leverage the power of higher degrees, we propose three architectural improvements -- attention re-normalization, separable \$S\textasciicircum 2\$ activation and separable layer normalization. Putting this all together, we propose EquiformerV2, which outperforms previous state-of-the-art methods on the large-scale OC20 dataset by up to \$12\textbackslash\%\$ on forces, \$4\textbackslash\%\$ on energies, offers better speed-accuracy trade-offs, and \$2\textbackslash times\$ reduction in DFT calculations needed for computing adsorption energies.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ablation study,AML,attention,Equiformer,equivariant,eSCN,GemNet,ML,MLP,OC20,Open Catalyst,QM9,SO(3),todo-tagging,transformer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Liao et al_2023_EquiformerV2.pdf;/Users/wasmer/Zotero/storage/MHXITTSP/2306.html}
}
@@ -9899,7 +10872,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
urldate = {2024-05-08},
abstract = {Equivariant Transformers such as Equiformer have demonstrated the efficacy of applying Transformers to the domain of 3D atomistic systems. However, they are limited to small degrees of equivariant representations due to their computational complexity. In this paper, we investigate whether these architectures can scale well to higher degrees. Starting from Equiformer, we first replace \$SO(3)\$ convolutions with eSCN convolutions to efficiently incorporate higher-degree tensors. Then, to better leverage the power of higher degrees, we propose three architectural improvements -- attention re-normalization, separable \$S\textasciicircum 2\$ activation and separable layer normalization. Putting this all together, we propose EquiformerV2, which outperforms previous state-of-the-art methods on large-scale OC20 dataset by up to \$9\textbackslash\%\$ on forces, \$4\textbackslash\%\$ on energies, offers better speed-accuracy trade-offs, and \$2\textbackslash times\$ reduction in DFT calculations needed for computing adsorption energies. Additionally, EquiformerV2 trained on only OC22 dataset outperforms GemNet-OC trained on both OC20 and OC22 datasets, achieving much better data efficiency. Finally, we compare EquiformerV2 with Equiformer on QM9 and OC20 S2EF-2M datasets to better understand the performance gain brought by higher degrees.},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ablation study,AML,attention,Equiformer,equivariant,eSCN,GemNet,ML,MLP,OC20,Open Catalyst,QM9,SO(3),transformer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Liao et al_2023_EquiformerV3.pdf}
}
@@ -9921,7 +10894,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
abstract = {Recent advances in machine learning (ML) have led to substantial performance improvement in material database benchmarks, but an excellent benchmark score may not imply good generalization performance. Here we show that ML models trained on Materials Project 2018 can have severely degraded performance on new compounds in Materials Project 2021 due to the distribution shift. We discuss how to foresee the issue with a few simple tools. Firstly, the uniform manifold approximation and projection (UMAP) can be used to investigate the relation between the training and test data within the feature space. Secondly, the disagreement between multiple ML models on the test data can illuminate out-of-distribution samples. We demonstrate that the UMAP-guided and query by committee acquisition strategies can greatly improve prediction accuracy by adding only 1\% of the test data. We believe this work provides valuable insights for building databases and models that enable better robustness and generalizability.},
issue = {1},
langid = {english},
- keywords = {\_tablet,ALIGNN,alloys,compositional descriptors,cross-validation,data imbalance,data scarcity,database generation,human bias,materials project,MP18,MP21,PCA,random forest,small data,train-test split,UMAP,unsupervised learning,with-code,with-data,XGB},
+ keywords = {ALIGNN,alloys,compositional descriptors,cross-validation,data imbalance,data scarcity,database generation,human bias,materials project,MP18,MP21,PCA,random forest,small data,train-test split,UMAP,unsupervised learning,with-code,with-data,XGB},
file = {/Users/wasmer/Nextcloud/Zotero/Li et al_2023_A critical examination of robustness and generalizability of machine learning.pdf}
}
@@ -9963,7 +10936,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
abstract = {The marriage of density functional theory (DFT) and deep-learning methods has the potential to revolutionize modern computational materials science. Here we develop a deep neural network approach to represent the DFT Hamiltonian (DeepH) of crystalline materials, aiming to bypass the computationally demanding self-consistent field iterations of DFT and substantially improve the efficiency of ab initio electronic-structure calculations. A general framework is proposed to deal with the large dimensionality and gauge (or rotation) covariance of the DFT Hamiltonian matrix by virtue of locality, and this is realized by a message-passing neural network for deep learning. High accuracy, high efficiency and good transferability of the DeepH method are generally demonstrated for various kinds of material system and physical property. The method provides a solution to the accuracy–efficiency dilemma of DFT and opens opportunities to explore large-scale material systems, as evidenced by a promising application in the study of twisted van der Waals materials.},
issue = {6},
langid = {english},
- keywords = {\_tablet,2D material,AML,Berry phase,CNT,DeepH,defects,DFT,disordered,e3nn,equivariant,GGA,graphene,heterostructures,library,local coordinates,materials,ML,ML-DFT,ML-ESM,MoS2,MPNN,near-sightedness,OpenMX,PBE,PCA,prediction of bandstructure,prediction of Berry phase,prediction of Hamiltonian matrix,SOC,twisted bilayer graphene,vdW,vdW materials,with-code,with-data},
+ keywords = {2D material,AML,Berry phase,CNT,DeepH,defects,DFT,disordered,e3nn,equivariant,GGA,graphene,heterostructures,library,local coordinates,materials,ML,ML-DFT,ML-ESM,MoS2,MPNN,near-sightedness,OpenMX,PBE,PCA,prediction of bandstructure,prediction of Berry phase,prediction of Hamiltonian matrix,SOC,twisted bilayer graphene,vdW,vdW materials,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Li et al_2022_Deep-learning density functional theory Hamiltonian for efficient ab initio.pdf}
}
@@ -10002,7 +10975,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
abstract = {Ab initio studies of magnetic superstructures are indispensable to research on emergent quantum materials, but are currently bottlenecked by the formidable computational cost. Here, to break this bottleneck, we have developed a deep equivariant neural network framework to represent the density functional theory Hamiltonian of magnetic materials for efficient electronic-structure calculation. A neural network architecture incorporating a priori knowledge of fundamental physical principles, especially the nearsightedness principle and the equivariance requirements of Euclidean and time-reversal symmetries (\$\$E(3)\textbackslash times \textbackslash\{I,\{\{\{\textbackslash mathcal\{T\}\}\}\}\textbackslash\}\$\$), is designed, which is critical to capture the subtle magnetic effects. Systematic experiments on spin-spiral, nanotube and moiré magnets were performed, making the challenging study of magnetic skyrmions feasible.},
issue = {4},
langid = {english},
- keywords = {\_tablet,AML,collinear,constrained DFT,DeepH,DFT,E(3),e3nn,ENN,equivariant,Heisenberg model,Jij,library,magnetic interactions,magnetic structure,magnetic supperlattice,magnetism,ML,ML-DFT,ML-ESM,MPNN,near-sightedness,non-collinear,OpenMX,PBE,prediction from magnetic configuration,prediction from structure,prediction of Hamiltonian matrix,prediction of Jij,prediction of magnetic moment,skyrmions,spin spiral,spin-dependent,transition metals,TRS,twisted bilayer,with-code,with-data},
+ keywords = {AML,collinear,constrained DFT,DeepH,DFT,E(3),e3nn,ENN,equivariant,Heisenberg model,Jij,library,magnetic interactions,magnetic structure,magnetic supperlattice,magnetism,ML,ML-DFT,ML-ESM,MPNN,near-sightedness,non-collinear,OpenMX,PBE,prediction from magnetic configuration,prediction from structure,prediction of Hamiltonian matrix,prediction of Jij,prediction of magnetic moment,skyrmions,spin spiral,spin-dependent,transition metals,TRS,twisted bilayer,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Li et al_2023_Deep-learning electronic-structure calculation of magnetic superstructures.pdf}
}
@@ -10011,12 +10984,12 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Li, He and Wang, Zun and Zou, Nianlong and Ye, Meng and Duan, Wenhui and Xu, Yong},
date = {2021-04-08},
eprint = {2104.03786},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, physics:quant-ph},
url = {http://arxiv.org/abs/2104.03786},
urldate = {2022-01-02},
abstract = {The marriage of density functional theory (DFT) and deep learning methods has the potential to revolutionize modern research of material science. Here we study the crucial problem of representing DFT Hamiltonian for crystalline materials of arbitrary configurations via deep neural network. A general framework is proposed to deal with the infinite dimensionality and covariance transformation of DFT Hamiltonian matrix in virtue of locality and use message passing neural network together with graph representation for deep learning. Our example study on graphene-based systems demonstrates that high accuracy (\$\textbackslash sim\$meV) and good transferability can be obtained for DFT Hamiltonian, ensuring accurate predictions of materials properties without DFT. The Deep Hamiltonian method provides a solution to the accuracy-efficiency dilemma of DFT and opens new opportunities to explore large-scale materials and physics.},
- keywords = {\_tablet,AML,DeepH,DFT,disordered,e3nn,equivariant,graphene,library,materials,ML,ML-DFT,ML-ESM,MPNN,OpenMX,original publication,PBE,prediction of Hamiltonian matrix,with-code},
+ keywords = {AML,DeepH,DFT,disordered,e3nn,equivariant,graphene,library,materials,ML,ML-DFT,ML-ESM,MPNN,OpenMX,original publication,PBE,prediction of Hamiltonian matrix,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Li et al_2021_Deep Neural Network Representation of Density Functional Theory Hamiltonian.pdf;/Users/wasmer/Zotero/storage/B7RUP7VH/2104.html}
}
@@ -10054,7 +11027,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
urldate = {2023-09-20},
abstract = {Rigorous expressions for the exchange parameters of classical Heisenberg model applied to crystals are obtained using a local spin density functional (LSDF) approach and KKR-Green functions formalism. The spin wave stiffness constant and Curie temperature (Tc) of ferromagnetic metals are obtained without any model assumptions as to the character of exchange interactions. The concentration dependence of Tc for binary ferromagnetic alloys is investigated in the framework of the single-site CPA-theory. The corresponding calculations are carried out for simple metals Fe, Ni and disordered Niî—¸Pd alloys.},
keywords = {alloys,calculation of Jij,CPA,Curie temperature,DFT,DFT theory,disordered,Ferromagnetism,infinitesimal rotation,Jij,KKR,KKR foundations,magnetism,original publication,physics},
- file = {/Users/wasmer/Zotero/storage/F46FQTHF/Liechtenstein et al. - 1987 - Local spin density functional approach to the theo.pdf;/Users/wasmer/Zotero/storage/23L5VB4T/0304885387907219.html}
+ file = {/Users/wasmer/Zotero/storage/F46FQTHF/Liechtenstein et al_1987_Local spin density functional approach to the theory of exchange interactions.pdf;/Users/wasmer/Zotero/storage/23L5VB4T/0304885387907219.html}
}
@article{liExploitingRedundancyLarge2023,
@@ -10093,7 +11066,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
urldate = {2024-04-18},
abstract = {The marriage of artificial intelligence (AI) and Materials Genome Initiative (MGI) could profoundly change the landscape of modern materials research, leading to a new paradigm of data-driven and AI-driven materials discovery. In this perspective, we will give an overview on the central role of AI in the MGI research. In particular, an emerging research field of ab initio AI, which applies state-of-the-art AI techniques to help solve bottleneck problems of ab initio computation, will be introduced. The development of ab initio AI will greatly accelerate high-throughput computation, promote the construction of large materials database, and open new opportunities for future research of MGI.},
langid = {english},
- keywords = {\_tablet,AI4Science,AML,autoencoder,DFT,diffusion model,DM21,Electronic structure,FermiNet,inverse design,materials discovery,Materials genome initiative,MD,ML,ML-DFA,ML-DFT,ML-ESM,perspective,QMC,review,VAE},
+ keywords = {AI4Science,AML,autoencoder,DFT,diffusion model,DM21,Electronic structure,FermiNet,inverse design,materials discovery,Materials genome initiative,MD,ML,ML-DFA,ML-DFT,ML-ESM,perspective,QMC,review,VAE},
file = {/Users/wasmer/Nextcloud/Zotero/Li et al_2023_Ab initio artificial intelligence.pdf;/Users/wasmer/Zotero/storage/SH6A3KT6/mgea.html}
}
@@ -10140,13 +11113,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lim, Derek and Robinson, Joshua and Zhao, Lingxiao and Smidt, Tess and Sra, Suvrit and Maron, Haggai and Jegelka, Stefanie},
date = {2022-09-30},
eprint = {2202.13013},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.2202.13013},
url = {http://arxiv.org/abs/2202.13013},
urldate = {2023-05-26},
abstract = {We introduce SignNet and BasisNet -- new neural architectures that are invariant to two key symmetries displayed by eigenvectors: (i) sign flips, since if \$v\$ is an eigenvector then so is \$-v\$; and (ii) more general basis symmetries, which occur in higher dimensional eigenspaces with infinitely many choices of basis eigenvectors. We prove that under certain conditions our networks are universal, i.e., they can approximate any continuous function of eigenvectors with the desired invariances. When used with Laplacian eigenvectors, our networks are provably more expressive than existing spectral methods on graphs; for instance, they subsume all spectral graph convolutions, certain spectral graph invariants, and previously proposed graph positional encodings as special cases. Experiments show that our networks significantly outperform existing baselines on molecular graph regression, learning expressive graph representations, and learning neural fields on triangle meshes. Our code is available at https://github.com/cptq/SignNet-BasisNet .},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,eigenvector,General ML,GNN,graph convolution,graph regression,library,ML,transformer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Lim et al_2022_Sign and Basis Invariant Networks for Spectral Graph Representation Learning.pdf;/Users/wasmer/Zotero/storage/JA5U2K7L/2202.html}
}
@@ -10170,13 +11143,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lin, Yuchao and Yan, Keqiang and Luo, Youzhi and Liu, Yi and Qian, Xiaoning and Ji, Shuiwang},
date = {2023-08-01},
eprint = {2306.10045},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2306.10045},
url = {http://arxiv.org/abs/2306.10045},
urldate = {2023-08-19},
abstract = {We study property prediction for crystal materials. A crystal structure consists of a minimal unit cell that is repeated infinitely in 3D space. How to accurately represent such repetitive structures in machine learning models remains unresolved. Current methods construct graphs by establishing edges only between nearby nodes, thereby failing to faithfully capture infinite repeating patterns and distant interatomic interactions. In this work, we propose several innovations to overcome these limitations. First, we propose to model physics-principled interatomic potentials directly instead of only using distances as in many existing methods. These potentials include the Coulomb potential, London dispersion potential, and Pauli repulsion potential. Second, we model the complete set of potentials among all atoms, instead of only between nearby atoms as in existing methods. This is enabled by our approximations of infinite potential summations with provable error bounds. We further develop efficient algorithms to compute the approximations. Finally, we propose to incorporate our computations of complete interatomic potentials into message passing neural networks for representation learning. We perform experiments on the JARVIS and Materials Project benchmarks for evaluation. Results show that the use of interatomic potentials and complete interatomic potentials leads to consistent performance improvements with reasonable computational costs. Our code is publicly available as part of the AIRS library (https://github.com/divelab/AIRS/tree/main/OpenMat/PotNet).},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Lin et al_2023_Efficient Approximations of Complete Interatomic Potentials for Crystal.pdf;/Users/wasmer/Zotero/storage/E3N59FIA/2306.html}
}
@@ -10186,14 +11159,14 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Li, Yang and Tang, Zechen and Chen, Zezhou and Sun, Minghui and Zhao, Boheng and Li, He and Tao, Honggeng and Yuan, Zilong and Duan, Wenhui and Xu, Yong},
date = {2024-03-17},
eprint = {2403.11287},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2403.11287},
url = {http://arxiv.org/abs/2403.11287},
urldate = {2024-04-18},
abstract = {Deep-learning density functional theory (DFT) shows great promise to significantly accelerate material discovery and potentially revolutionize materials research, which demands a close combination between neural networks and DFT computation. However, current research in this field primarily relies on supervised learning, making the developments of neural networks and DFT isolated from each other. In this work, we present a theoretical framework of neural-network DFT, which unifies the optimization of neural networks with the variational computation of DFT, enabling physics-informed unsupervised learning. Moreover, we develop a differential DFT code incorporated with deep-learning DFT Hamiltonian, and introduce algorithms of automatic differentiation and backpropagation to DFT, demonstrating the concept of neural-network DFT. The advanced neural-network architecture not only surpasses conventional approaches in accuracy and efficiency, but offers a new paradigm for developing deep-learning DFT methods.},
- pubstate = {preprint},
- keywords = {/unread,\_tablet,Condensed Matter - Materials Science,Physics - Computational Physics},
+ pubstate = {prepublished},
+ keywords = {/unread,Condensed Matter - Materials Science,Physics - Computational Physics},
file = {/Users/wasmer/Nextcloud/Zotero/Li et al_2024_Neural-network density functional theory.pdf;/Users/wasmer/Zotero/storage/J4JJX4HR/2403.html}
}
@@ -10238,13 +11211,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lipton, Zachary C. and Steinhardt, Jacob},
date = {2018-07-26},
eprint = {1807.03341},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.1807.03341},
url = {http://arxiv.org/abs/1807.03341},
urldate = {2022-06-27},
abstract = {Collectively, machine learning (ML) researchers are engaged in the creation and dissemination of knowledge about data-driven algorithms. In a given paper, researchers might aspire to any subset of the following goals, among others: to theoretically characterize what is learnable, to obtain understanding through empirically rigorous experiments, or to build a working system that has high predictive accuracy. While determining which knowledge warrants inquiry may be subjective, once the topic is fixed, papers are most valuable to the community when they act in service of the reader, creating foundational knowledge and communicating as clearly as possible. Recent progress in machine learning comes despite frequent departures from these ideals. In this paper, we focus on the following four patterns that appear to us to be trending in ML scholarship: (i) failure to distinguish between explanation and speculation; (ii) failure to identify the sources of empirical gains, e.g., emphasizing unnecessary modifications to neural architectures when gains actually stem from hyper-parameter tuning; (iii) mathiness: the use of mathematics that obfuscates or impresses rather than clarifies, e.g., by confusing technical and non-technical concepts; and (iv) misuse of language, e.g., by choosing terms of art with colloquial connotations or by overloading established technical terms. While the causes behind these patterns are uncertain, possibilities include the rapid expansion of the community, the consequent thinness of the reviewer pool, and the often-misaligned incentives between scholarship and short-term measures of success (e.g., bibliometrics, attention, and entrepreneurial opportunity). While each pattern offers a corresponding remedy (don't do it), we also discuss some speculative suggestions for how the community might combat these trends.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {criticism,ML,research ethics,skepticism,state of a field},
file = {/Users/wasmer/Nextcloud/Zotero/Lipton_Steinhardt_2018_Troubling Trends in Machine Learning Scholarship.pdf;/Users/wasmer/Zotero/storage/HK89ZR8C/1807.html}
}
@@ -10254,13 +11227,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Li, Xihan and Chen, Xiang and Tutunov, Rasul and Bou-Ammar, Haitham and Wang, Lei and Wang, Jun},
date = {2022-02-02},
eprint = {2202.01388},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {quant-ph},
doi = {10.48550/arXiv.2202.01388},
url = {http://arxiv.org/abs/2202.01388},
urldate = {2023-09-23},
abstract = {The Schr\textbackslash "odinger equation is at the heart of modern quantum mechanics. Since exact solutions of the ground state are typically intractable, standard approaches approximate Schr\textbackslash "odinger equation as forms of nonlinear generalized eigenvalue problems \$F(V)V = SV\textbackslash Lambda\$ in which \$F(V)\$, the matrix to be decomposed, is a function of its own top-\$k\$ smallest eigenvectors \$V\$, leading to a "self-consistency problem". Traditional iterative methods heavily rely on high-quality initial guesses of \$V\$ generated via domain-specific heuristics methods based on quantum mechanics. In this work, we eliminate such a need for domain-specific heuristics by presenting a novel framework, Self-consistent Gradient-like Eigen Decomposition (SCGLED) that regards \$F(V)\$ as a special "online data generator", thus allows gradient-like eigendecomposition methods in streaming \$k\$-PCA to approach the self-consistency of the equation from scratch in an iterative way similar to online learning. With several critical numerical improvements, SCGLED is robust to initial guesses, free of quantum-mechanism-based heuristics designs, and neat in implementation. Our experiments show that it not only can simply replace traditional heuristics-based initial guess methods with large performance advantage (achieved averagely 25x more precise than the best baseline in similar wall time), but also is capable of finding highly precise solutions independently without any traditional iterative methods.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,alternative approaches,eigendecomposition,generalized eigenvalue problem,incremental algorithms,iterative algorithms,numerical linear algebra,online learning,optimization,SCF,Schrödinger equation,streaming algorithm,streaming algorithms},
file = {/Users/wasmer/Zotero/storage/DHLPNXT3/Li et al. - 2022 - Self-consistent Gradient-like Eigen Decomposition .pdf;/Users/wasmer/Zotero/storage/TAQH9FC5/2202.html}
}
@@ -10294,7 +11267,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
url = {http://jmlr.org/papers/v22/21-0343.html},
urldate = {2023-10-13},
abstract = {Although there exist several libraries for deep learning on graphs, they are aiming at implementing basic operations for graph deep learning. In the research community, implementing and benchmarking various advanced tasks are still painful and time-consuming with existing libraries. To facilitate graph deep learning research, we introduce DIG: Dive into Graphs, a turnkey library that provides a unified testbed for higher level, research-oriented graph deep learning tasks. Currently, we consider graph generation, self-supervised learning on graphs, explainability of graph neural networks, and deep learning on 3D graphs. For each direction, we provide unified implementations of data interfaces, common algorithms, and evaluation metrics. Altogether, DIG is an extensible, open-source, and turnkey library for researchers to develop new methods and effortlessly compare with common baselines using widely used datasets and evaluation metrics. Source code is available at https://github.com/divelab/DIG.},
- keywords = {\_tablet,alternative approaches,AML,DimeNet,General ML,geometric deep learning,GNN,GNNExplainer,graph generation,graph ML,library,LIFT,MD17,ML,MPNN,PyG,PyTorch,QM9,SchNet,SphereNet,SSL,with-code,XAI},
+ keywords = {alternative approaches,AML,DimeNet,General ML,geometric deep learning,GNN,GNNExplainer,graph generation,graph ML,library,LIFT,MD17,ML,MPNN,PyG,PyTorch,QM9,SchNet,SphereNet,SSL,with-code,XAI},
file = {/Users/wasmer/Nextcloud/Zotero/Liu et al_2021_DIG.pdf;/Users/wasmer/Zotero/storage/P9LJWCS3/DIG.html}
}
@@ -10361,13 +11334,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Liu, Ziming and Wang, Yixuan and Vaidya, Sachin and Ruehle, Fabian and Halverson, James and SoljaÄić, Marin and Hou, Thomas Y. and Tegmark, Max},
date = {2024-05-02},
eprint = {2404.19756},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, stat},
doi = {10.48550/arXiv.2404.19756},
url = {http://arxiv.org/abs/2404.19756},
urldate = {2024-05-05},
abstract = {Inspired by the Kolmogorov-Arnold representation theorem, we propose Kolmogorov-Arnold Networks (KANs) as promising alternatives to Multi-Layer Perceptrons (MLPs). While MLPs have fixed activation functions on nodes ("neurons"), KANs have learnable activation functions on edges ("weights"). KANs have no linear weights at all -- every weight parameter is replaced by a univariate function parametrized as a spline. We show that this seemingly simple change makes KANs outperform MLPs in terms of accuracy and interpretability. For accuracy, much smaller KANs can achieve comparable or better accuracy than much larger MLPs in data fitting and PDE solving. Theoretically and empirically, KANs possess faster neural scaling laws than MLPs. For interpretability, KANs can be intuitively visualized and can easily interact with human users. Through two examples in mathematics and physics, KANs are shown to be useful collaborators helping scientists (re)discover mathematical and physical laws. In summary, KANs are promising alternatives for MLPs, opening opportunities for further improving today's deep learning models which rely heavily on MLPs.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AI4Science,alternative approaches,alternative to MLP,Deep learning,General ML,hot topic,KAN,ML,original publication,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Liu et al_2024_KAN.pdf;/Users/wasmer/Zotero/storage/TUFV5YR7/2404.html}
}
@@ -10391,13 +11364,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Liu, Xianglin and Zhang, Jiaxin and Eisenbach, Markus and Wang, Yang},
date = {2019-12-31},
eprint = {1912.13460},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.1912.13460},
url = {http://arxiv.org/abs/1912.13460},
urldate = {2023-09-19},
abstract = {The development of machine learning sheds new light on the problem of statistical thermodynamics in multicomponent alloys. However, a data-driven approach to construct the effective Hamiltonian requires sufficiently large data sets, which is expensive to calculate with conventional density functional theory (DFT). To solve this problem, we propose to use the atomic local energy as the target variable, and harness the power of the linear-scaling DFT to accelerate the data generating process. Using the large amounts of DFT data sets, various complex models are devised and applied to learn the effective Hamiltonians of a range of refractory high entropy alloys (HEAs). The testing \$R\textasciicircum 2\$ scores of the effective pair interaction model are higher than 0.99, demonstrating that the pair interactions within the 6-th coordination shell provide an excellent description of the atomic local energies for all the four HEAs. This model is further improved by including nonlinear and multi-site interactions. In particular, the deep neural networks (DNNs) built directly in the local configuration space (therefore no hand-crafted features) are employed to model the effective Hamiltonian. The results demonstrate that neural networks are promising for the modeling of effective Hamiltonian due to its excellent representation power.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Liu et al_2019_Machine Learning the Effective Hamiltonian in High Entropy Alloys.pdf;/Users/wasmer/Zotero/storage/GU92DRKM/1912.html}
}
@@ -10426,13 +11399,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Liu, Xianglin and Zhang, Jiaxin and Eisenbach, Markus and Wang, Yang},
date = {2019-06-07},
eprint = {1906.02889},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.1906.02889},
url = {http://arxiv.org/abs/1906.02889},
urldate = {2023-09-19},
abstract = {The development of machine learning sheds new light on the traditionally complicated problem of thermodynamics in multicomponent alloys. Successful application of such a method, however, strongly depends on the quality of the data and model. Here we propose a scheme to improve the representativeness of the data by utilizing the short-range order (SRO) parameters to survey the configuration space. Using the improved data, a pair interaction model is trained for the NbMoTaW high entropy alloy using linear regression. Benefiting from the physics incorporated into the model, the learned effective Hamiltonian demonstrates excellent predictability over the whole configuration space. By including pair interactions within the 6th nearest-neighbor shell, this model achieves an \$R\textasciicircum 2\$ testing score of 0.997 and root mean square error of 0.43 meV. We further perform a detailed analysis on the effects of training data, testing data, and model parameters. The results reveal the vital importance of representative data and physical model. On the other hand, we also examined the performance neural networks, which is found to demonstrate a strong tendency to overfit the data.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Liu et al_2019_Machine learning modeling of high entropy alloy.pdf;/Users/wasmer/Zotero/storage/RIJJA86L/1906.html}
}
@@ -10470,7 +11443,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
urldate = {2023-06-15},
abstract = {Topological insulators (TIs) provide intriguing prospects for the future of spintronics due to their large spin–orbit coupling and dissipationless, counter-propagating conduction channels in the surface state. The combination of topological properties and magnetic order can lead to new quantum states including the quantum anomalous Hall effect that was first experimentally realized in Cr-doped (Bi,Sb)2Te3 films. Since magnetic doping can introduce detrimental effects, requiring very low operational temperatures, alternative approaches are explored. Proximity coupling to magnetically ordered systems is an obvious option, with the prospect to raise the temperature for observing the various quantum effects. Here, an overview of proximity coupling and interfacial effects in TI heterostructures is presented, which provides a versatile materials platform for tuning the magnetic and topological properties of these exciting materials. An introduction is first given to the heterostructure growth by molecular beam epitaxy and suitable structural, electronic, and magnetic characterization techniques. Going beyond transition-metal-doped and undoped TI heterostructures, examples of heterostructures are discussed, including rare-earth-doped TIs, magnetic insulators, and antiferromagnets, which lead to exotic phenomena such as skyrmions and exchange bias. Finally, an outlook on novel heterostructures such as intrinsic magnetic TIs and systems including 2D materials is given.},
langid = {english},
- keywords = {/unread,\_tablet,Hall effect,Hall QAHE,magnetic TIs,review,topological insulator},
+ keywords = {/unread,Hall effect,Hall QAHE,magnetic TIs,review,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Liu_Hesjedal_2021_Magnetic Topological Insulator Heterostructures.pdf;/Users/wasmer/Zotero/storage/VEP2MG97/adma.html}
}
@@ -10488,7 +11461,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
urldate = {2022-07-08},
abstract = {Machine learning (ML) is an increasingly popular statistical tool for analyzing either measured or calculated data sets. Here, we explore its application to a well-defined physics problem, investigating issues of how the underlying physics is handled by ML, and how self-consistent solutions can be found by limiting the domain in which ML is applied. The particular problem is how to find accurate approximate density functionals for the kinetic energy (KE) of noninteracting electrons. Kernel ridge regression is used to approximate the KE of non-interacting fermions in a one dimensional box as a functional of their density. The properties of different kernels and methods of cross-validation are explored, reproducing the physics faithfully in some cases, but not others. We also address how self-consistency can be achieved with information on only a limited electronic density domain. Accurate constrained optimal densities are found via a modified Euler-Lagrange constrained minimization of the machine-learned total energy, despite the poor quality of its functional derivative. A projected gradient descent algorithm is derived using local principal component analysis. Additionally, a sparse grid representation of the density can be used without degrading the performance of the methods. The implications for machine-learned density functional approximations are discussed. © 2015 Wiley Periodicals, Inc.},
langid = {english},
- keywords = {\_tablet,DFA,DFT,KRR,ML,ML-DFA,ML-DFT,ML-ESM,tutorial},
+ keywords = {DFA,DFT,KRR,ML,ML-DFA,ML-DFT,ML-ESM,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Li et al_2016_Understanding machine-learned density functionals.pdf;/Users/wasmer/Zotero/storage/ZPNDJ7AU/qua.html}
}
@@ -10577,7 +11550,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
url = {https://link.aps.org/doi/10.1103/PhysRevMaterials.7.045802},
urldate = {2023-05-05},
abstract = {Alloys composed of several elements in roughly equimolar composition, often referred to as high-entropy alloys, have long been of interest for their thermodynamics and peculiar mechanical properties, and more recently for their potential application in catalysis. They are a considerable challenge to traditional atomistic modeling, and also to data-driven potentials that for the most part have memory footprint, computational effort, and data requirements which scale poorly with the number of elements included. We apply a recently proposed scheme to compress chemical information in a lower-dimensional space, which reduces dramatically the cost of the model with negligible loss of accuracy, to build a potential that can describe 25 d-block transition metals. The model shows semiquantitative accuracy for prototypical alloys and is remarkably stable when extrapolating to structures outside its training set. We use this framework to study element segregation in a computational experiment that simulates an equimolar alloy of all 25 elements, mimicking the seminal experiments in the groups of Yeh and Cantor, and use our observations on the short-range order relations between the elements to define a data-driven set of Hume-Rothery rules that can serve as guidance for alloy design. We conclude with a study of three prototypical alloys, CoCrFeMnNi, CoCrFeMoNi, and IrPdPtRhRu, determining their stability and the short-range order behavior of their constituents.},
- keywords = {\_tablet,ACE,alchemical,chemical species scaling problem,descriptors,dimensionality reduction,high-entropy alloys,library,MTP,PyTorch,SOAP,transition metals,with-code},
+ keywords = {ACE,alchemical,chemical species scaling problem,descriptors,dimensionality reduction,high-entropy alloys,library,MTP,PyTorch,SOAP,transition metals,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Lopanitsyna et al_2023_Modeling high-entropy transition metal alloys with alchemical compression.pdf;/Users/wasmer/Zotero/storage/WWWS3KMP/PhysRevMaterials.7.html}
}
@@ -10586,14 +11559,14 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lopanitsyna, Nataliya and Fraux, Guillaume and Springer, Maximilian A. and De, Sandip and Ceriotti, Michele},
date = {2022-12-26},
eprint = {2212.13254},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2212.13254},
url = {http://arxiv.org/abs/2212.13254},
urldate = {2022-12-29},
abstract = {Alloys composed of several elements in roughly equimolar composition, often referred to as high-entropy alloys, have long been of interest for their thermodynamics and peculiar mechanical properties, and more recently for their potential application in catalysis. They are a considerable challenge to traditional atomistic modeling, and also to data-driven potentials that for the most part have memory footprint, computational effort and data requirements which scale poorly with the number of elements included. We apply a recently proposed scheme to compress chemical information in a lower-dimensional space, which reduces dramatically the cost of the model with negligible loss of accuracy, to build a potential that can describe 25 d-block transition metals. The model shows semi-quantitative accuracy for prototypical alloys, and is remarkably stable when extrapolating to structures outside its training set. We use this framework to study element segregation in a computational experiment that simulates an equimolar alloy of all 25 elements, mimicking the seminal experiments by Cantor et al., and use our observations on the short-range order relations between the elements to define a data-driven set of Hume-Rothery rules that can serve as guidance for alloy design. We conclude with a study of three prototypical alloys, CoCrFeMnNi, CoCrFeMoNi and IrPdPtRhRu, determining their stability and the short-range order behavior of their constituents.},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,alchemical,chemical species scaling problem,descriptors,dimensionality reduction,high-entropy alloys,MTP,PyTorch,SOAP,transition metals,with-code},
+ pubstate = {prepublished},
+ keywords = {ACE,alchemical,chemical species scaling problem,descriptors,dimensionality reduction,high-entropy alloys,MTP,PyTorch,SOAP,transition metals,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Lopanitsyna et al_2022_Modeling high-entropy transition-metal alloys with alchemical compression.pdf;/Users/wasmer/Nextcloud/Zotero/Lopanitsyna et al_2022_Modeling high-entropy transition-metal alloys with alchemical compression2.pdf;/Users/wasmer/Zotero/storage/QNGQ9AQD/2212.html}
}
@@ -10606,7 +11579,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
abstract = {The question how magnetism behaves when the dimension of materials is reduced to increasingly smaller sizes has attracted much research and led to the development of the field of magnetic nanostructures. This research has been fueled by the technological potential of these systems for the field of high-density magnetic storage media and has been accelerated by the many novel experimental methods and techniques developed exhibiting atomic resolution. This thesis is motivated by the quest for the understanding and the exploration of complex magnetism provided by atomic scale magnetic clusters deposited on surfaces or embedded in the bulk. The nature of magnetism in these systems can be very rich, in that the properties depend on the atomic species, the cluster size, shape and symmetry or choice of the substrate. Small variations of the cluster parameter may change the properties dramatically. Particularly rich and particularly challenging for experiment and theory is the behavior of clusters with competing magnetic interactions either between the cluster atoms or between the cluster and the substrate. In both cases magnetic frustration can lead to non-collinear magnetic structures for which the magnetic quantization axis changes from atom to atom. This thesis sheds light onto these systems from a theoretical perspective. Use is made of the density functional theory (DFT), the most successful material specific theory for describing electronic and derived properties from first-principles. Acting within this framework, we have developed and implemented the treatment of non-collinear magnetism into the Jülich version of the full-potential Korringa-Kohn-Rostoker Green Function (KKR-GF) method. The KKR-GF method provides several advantages compared to other first-principles methods. Based on solving the Dyson equation it allows an elegant treatment of non-periodic systems such as impurities and clusters in bulk or on surfaces. Electronic, magnetic properties and the observables provided by experimental techniques such as x-ray, scanning tunneling microscopy and spectroscopy can be accessed with the KKR-GF method. Firstly, the method was applied to 3\$\textbackslash textit\{d\}\$ transition-metal clusters on different ferromagnetic surfaces. Different types of magnetic clusters where selected. Clusters of Fe, Co, Ni atoms are ferromagnetic and thus magnetically collinear. In order to investigate magnetic frustration due to competing interactions within the ad-cluster we considered a (001) oriented surface of \$\textbackslash textit\{fcc\}\$ metals, a topology which usually does not lead to non-collinear magnetism. We tuned the strength of the magnetic coupling between the ad-clusters and the ferromagnetic surface by varying the substrate from the case of Ni(001) with a rather weak hybridization of the Ni \$\textbackslash textit\{d\}\$-states with the adatom \$\textbackslash textit\{d\}\$-states to the case of Fe\$\_\{3ML\}\$/Cu(001) with a much stronger hybridization due to the larger extend of the Fe wavefunctions. On Ni(001), the interaction between the Cr- as well as the Mn-dimer adatoms is of antiferromagnetic nature, which is in competition with the interaction with the substrate atoms. If we allow the magnetism to be non-collinear, the moments rotate such the Cr-(Mn) adatom moments are aligned antiparallel to each other and are basically perpendicular to the substrate moments. However, the weak AF(FM) interaction with the substrate causes a slight tilting towards the substrate, leading to an angle of 94.2◦(72.6◦) instead of 90◦. After performing total energy calculations we find that for Cr-dimer the ground state is collinear whereas the Mn-dimer prefers the non-collinear configuration as ground state. The Heisenberg model is shown [...]},
langid = {english},
pagetotal = {189},
- keywords = {\_tablet,Dissertation (Univ.),Hochschulschrift,juKKR,KKR,magnetism,PGI-1/IAS-1,thesis},
+ keywords = {Dissertation (Univ.),Hochschulschrift,juKKR,KKR,magnetism,PGI-1/IAS-1,thesis},
file = {/Users/wasmer/Nextcloud/Zotero/Lounis_2007_Theory of Magnetic Transition Metal Nanoclusters on Surfaces.pdf}
}
@@ -10667,7 +11640,7 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
abstract = {Having served as a playground for fundamental studies on the physics of d and f electrons for almost a century, magnetic molecules are now becoming increasingly important for technological applications, such as magnetic resonance, data storage, spintronics and quantum information. All of these applications require the preservation and control of spins in time, an ability hampered by the interaction with the environment, namely with other spins, conduction electrons, molecular vibrations and electromagnetic fields. Thus, the design of a novel magnetic molecule with tailored properties is a formidable task, which does not only concern its electronic structures but also calls for a deep understanding of the interaction among all the degrees of freedom at play. This Review describes how state-of-the-art ab initio computational methods, combined with data-driven approaches to materials modelling, can be integrated into a fully multiscale strategy capable of defining design rules for magnetic molecules.},
issue = {11},
langid = {english},
- keywords = {Computational chemistry,Electronic structure,Magnetic materials,Magnetic properties and materials,Spintronics},
+ keywords = {\_tablet,Computational chemistry,Electronic structure,Magnetic materials,Magnetic properties and materials,Spintronics},
file = {/Users/wasmer/Nextcloud/Zotero/Lunghi_Sanvito_2022_Computational design of magnetic molecules and their environment using quantum.pdf}
}
@@ -10677,17 +11650,35 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Lunghi, Alessandro and Sanvito, Stefano},
date = {2019-11-06},
eprint = {1911.02263},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.1911.02263},
url = {http://arxiv.org/abs/1911.02263},
urldate = {2023-02-23},
abstract = {The advent of computational statistical disciplines, such as machine learning, is leading to a paradigm shift in the way we conceive the design of new compounds. Today computational science does not only provide a sound understanding of experiments, but also can directly design the best compound for specific applications. This approach, known as reverse engineering, requires the construction of models able to efficiently predict continuous structure-property maps. Here we show that reverse engineering can be used to tune the magnetic properties of a single-ion molecular magnet in an automated intelligent fashion. We design a machine learning model to predict both the energy and magnetic properties as function of the chemical structure. Then, a particle-swarm optimization algorithm is used to explore the conformational landscapes in the search for new molecular structures leading to an enhanced magnetic anisotropy. We find that a 5\% change in one of the coordination angles leads to a 50\% increase in the anisotropy. Our approach paves the way for a machine-learning-driven exploration of the chemical space of general classes of magnetic materials. Most importantly, it can be applied to any structure-property relation and offers an effective way to automatically generate new materials with target properties starting from the knowledge of previously synthesized ones.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,\_tablet,Condensed Matter - Materials Science,Physics - Computational Physics},
file = {/Users/wasmer/Nextcloud/Zotero/Lunghi_Sanvito_2019_Surfing multiple conformation-property landscapes via machine learning.pdf;/Users/wasmer/Zotero/storage/FQSQYUBP/1911.html}
}
+@article{luoDataDrivenCompressionElectronPhonon2024,
+ title = {Data-{{Driven Compression}} of {{Electron-Phonon Interactions}}},
+ author = {Luo, Yao and Desai, Dhruv and Chang, Benjamin K. and Park, Jinsoo and Bernardi, Marco},
+ date = {2024-05-01},
+ journaltitle = {Physical Review X},
+ shortjournal = {Phys. Rev. X},
+ volume = {14},
+ number = {2},
+ pages = {021023},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevX.14.021023},
+ url = {https://link.aps.org/doi/10.1103/PhysRevX.14.021023},
+ urldate = {2024-05-29},
+ abstract = {First-principles calculations of electron interactions in materials have seen rapid progress in recent years, with electron-phonon (ð‘’−ph) interactions being a prime example. However, these techniques use large matrices encoding the interactions on dense momentum grids, which reduces computational efficiency and obscures interpretability. For ð‘’−ph interactions, existing interpolation techniques leverage locality in real space, but the high dimensionality of the data remains a bottleneck to balance cost and accuracy. Here we show an efficient way to compress ð‘’−ph interactions based on singular value decomposition (SVD), a widely used matrix and image compression technique. Leveraging (un)constrained SVD methods, we accurately predict material properties related to ð‘’−ph interactions—including charge mobility, spin relaxation times, band renormalization, and superconducting critical temperature—while using only a small fraction (1\%–2\%) of the interaction data. These findings unveil the hidden low-dimensional nature of ð‘’−ph interactions. Furthermore, they accelerate state-of-the-art first-principles ð‘’−ph calculations by about 2 orders of magnitude without sacrificing accuracy. Our Pareto-optimal parametrization of ð‘’−ph interactions can be readily generalized to electron-electron and electron-defect interactions, as well as to other couplings, advancing quantitative studies of condensed matter.},
+ keywords = {/unread,alternative approaches,alternative for ML-DFT,DFPT,DFT,electron-phonon interaction,numerical linear algebra,phonon,scattering theory,speedup,superconductor,SVD},
+ file = {/Users/wasmer/Nextcloud/Zotero/Luo et al_2024_Data-Driven Compression of Electron-Phonon Interactions.pdf;/Users/wasmer/Zotero/storage/RI7JGWLJ/PhysRevX.14.html}
+}
+
@inproceedings{lupopasiniFastAccuratePredictions2022,
title = {Fast and {{Accurate Predictions}} of {{Total Energy}} for {{Solid Solution Alloys}} with {{Graph Convolutional Neural Networks}}},
booktitle = {Driving {{Scientific}} and {{Engineering Discoveries Through}} the {{Integration}} of {{Experiment}}, {{Big Data}}, and {{Modeling}} and {{Simulation}}},
@@ -10752,19 +11743,38 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
file = {/Users/wasmer/Nextcloud/Zotero/Lupo Pasini et al_2021_A scalable algorithm for the optimization of neural network architectures.pdf}
}
+@article{lvDeepChargeDeep2023,
+ title = {Deep {{Charge}}: {{Deep}} Learning Model of Electron Density from a One-Shot Density Functional Theory Calculation},
+ shorttitle = {Deep {{Charge}}},
+ author = {Lv, Taoyuze and Zhong, Zhicheng and Liang, Yuhang and Li, Feng and Huang, Jun and Zheng, Rongkun},
+ date = {2023-12-20},
+ journaltitle = {Physical Review B},
+ shortjournal = {Phys. Rev. B},
+ volume = {108},
+ number = {23},
+ pages = {235159},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevB.108.235159},
+ url = {https://link.aps.org/doi/10.1103/PhysRevB.108.235159},
+ urldate = {2024-05-27},
+ abstract = {Electron charge density is a fundamental physical quantity, determining various properties of matter. In this study, we have proposed a deep learning model for accurate charge-density prediction. Our model naturally preserves physical symmetries and can be effectively trained from one-shot density functional theory calculation toward high accuracy. It captures detailed atomic environment information, ensuring accurate predictions of charge density across bulk, surface, molecules, and amorphous structures. This implementation exhibits excellent scalability and provides efficient analyses of material properties in large-scale condensed matter systems.},
+ keywords = {alloys,AML,binary systems,disordered,grid-based descriptors,ML,ML-Density,ML-DFT,ML-ESM,prediction of electron density,silicon},
+ file = {/Users/wasmer/Nextcloud/Zotero/Lv et al_2023_Deep Charge.pdf;/Users/wasmer/Zotero/storage/2ZZKFF53/PhysRevB.108.html}
+}
+
@online{lysogorskiyActiveLearningStrategies2022,
title = {Active Learning Strategies for Atomic Cluster Expansion Models},
author = {Lysogorskiy, Yury and Bochkarev, Anton and Mrovec, Matous and Drautz, Ralf},
date = {2022-12-16},
eprint = {2212.08716},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2212.08716},
url = {http://arxiv.org/abs/2212.08716},
urldate = {2023-01-20},
abstract = {The atomic cluster expansion (ACE) was proposed recently as a new class of data-driven interatomic potentials with a formally complete basis set. Since the development of any interatomic potential requires a careful selection of training data and thorough validation, an automation of the construction of the training dataset as well as an indication of a model's uncertainty are highly desirable. In this work, we compare the performance of two approaches for uncertainty indication of ACE models based on the D-optimality criterion and ensemble learning. While both approaches show comparable predictions, the extrapolation grade based on the D-optimality (MaxVol algorithm) is more computationally efficient. In addition, the extrapolation grade indicator enables an active exploration of new structures, opening the way to the automated discovery of rare-event configurations. We demonstrate that active learning is also applicable to explore local atomic environments from large-scale MD simulations.},
- pubstate = {preprint},
- keywords = {ACE,active learning,Condensed Matter - Materials Science,D-optimality,database generation],descriptors,ensemble learning,iterative learning,iterative learning scheme,MD,molecular dynamics,uncertainty quantification},
+ pubstate = {prepublished},
+ keywords = {ACE,active learning,Condensed Matter - Materials Science,D-optimality,database generation,descriptors,ensemble learning,iterative learning,iterative learning scheme,MD,molecular dynamics,uncertainty quantification},
file = {/Users/wasmer/Nextcloud/Zotero/Lysogorskiy et al_2022_Active learning strategies for atomic cluster expansion models.pdf;/Users/wasmer/Zotero/storage/67ZIBP4V/2212.html}
}
@@ -10785,10 +11795,28 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
abstract = {The atomic cluster expansion is a general polynomial expansion of the atomic energy in multi-atom basis functions. Here we implement the atomic cluster expansion in the performant C++ code PACE that is suitable for use in large-scale atomistic simulations. We briefly review the atomic cluster expansion and give detailed expressions for energies and forces as well as efficient algorithms for their evaluation. We demonstrate that the atomic cluster expansion as implemented in PACE shifts a previously established Pareto front for machine learning interatomic potentials toward faster and more accurate calculations. Moreover, general purpose parameterizations are presented for copper and silicon and evaluated in detail. We show that the Cu and Si potentials significantly improve on the best available potentials for highly accurate large-scale atomistic simulations.},
issue = {1},
langid = {english},
- keywords = {\_tablet,ACE,C++,descriptors,library},
+ keywords = {ACE,C++,descriptors,library},
file = {/Users/wasmer/Nextcloud/Zotero/Lysogorskiy et al_2021_Performant implementation of the atomic cluster expansion (PACE) and.pdf;/Users/wasmer/Zotero/storage/QVQD97QT/s41524-021-00559-9.html}
}
+@article{m.casaresGradDFTSoftwareLibrary2024,
+ title = {{{GradDFT}}. {{A}} Software Library for Machine Learning Enhanced Density Functional Theory},
+ author = {M. Casares, Pablo A. and Baker, Jack S. and Medvidović, Matija and family=Reis, given=Roberto, prefix=dos, useprefix=false and Arrazola, Juan Miguel},
+ date = {2024-02-13},
+ journaltitle = {The Journal of Chemical Physics},
+ shortjournal = {The Journal of Chemical Physics},
+ volume = {160},
+ number = {6},
+ pages = {062501},
+ issn = {0021-9606},
+ doi = {10.1063/5.0181037},
+ url = {https://doi.org/10.1063/5.0181037},
+ urldate = {2024-05-28},
+ abstract = {Density functional theory (DFT) stands as a cornerstone method in computational quantum chemistry and materials science due to its remarkable versatility and scalability. Yet, it suffers from limitations in accuracy, particularly when dealing with strongly correlated systems. To address these shortcomings, recent work has begun to explore how machine learning can expand the capabilities of DFT: an endeavor with many open questions and technical challenges. In this work, we present GradDFT a fully differentiable JAX-based DFT library, enabling quick prototyping and experimentation with machine learning-enhanced exchange–correlation energy functionals. GradDFT employs a pioneering parametrization of exchange–correlation functionals constructed using a weighted sum of energy densities, where the weights are determined using neural networks. Moreover, GradDFT encompasses a comprehensive suite of auxiliary functions, notably featuring a just-in-time compilable and fully differentiable self-consistent iterative procedure. To support training and benchmarking efforts, we additionally compile a curated dataset of experimental dissociation energies of dimers, half of which contain transition metal atoms characterized by strong electronic correlations. The software library is tested against experimental results to study the generalization capabilities of a neural functional across potential energy surfaces and atomic species, as well as the effect of training data noise on the resulting model accuracy.},
+ keywords = {/unread,AML,autodiff,DM21,JAX,library,ML,ML-DFA,ML-DFT,ML-ESM,prediction of Exc,transition metals,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/M. Casares et al_2024_GradDFT.pdf;/Users/wasmer/Zotero/storage/7P6AHS6Z/GradDFT-A-software-library-for-machine-learning.html}
+}
+
@article{m.tealeDFTExchangeSharing2022,
title = {{{DFT}} Exchange: Sharing Perspectives on the Workhorse of Quantum Chemistry and Materials Science},
shorttitle = {{{DFT}} Exchange},
@@ -10859,13 +11887,13 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
author = {Mandal, Manasi and Drucker, Nathan C. and Siriviboon, Phum and Nguyen, Thanh and Boonkird, Tiya and Lamichhane, Tej Nath and Okabe, Ryotaro and Chotrattanapituk, Abhijatmedhi and Li, Mingda},
date = {2023-03-27},
eprint = {2303.15581},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2303.15581},
url = {http://arxiv.org/abs/2303.15581},
urldate = {2023-06-12},
abstract = {Topological superconductors (TSCs) have garnered significant research and industry attention in the past two decades. By hosting Majorana bound states which can be used as qubits that are robust against local perturbations, TSCs offer a promising platform toward (non-universal) topological quantum computation. However, there has been a scarcity of TSC candidates, and the experimental signatures that identify a TSC are often elusive. In this perspective, after a short review of the TSC basics and theories, we provide an overview of the TSC materials candidates, including natural compounds and synthetic material systems. We further introduce various experimental techniques to probe TSC, focusing on how a system is identified as a TSC candidate, and why a conclusive answer is often challenging to draw. We conclude by calling for new experimental signatures and stronger computational support to accelerate the search for new TSC candidates.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {materials,perspective,physics,superconductor,Topological Superconductor},
file = {/Users/wasmer/Zotero/storage/4J3A87J3/Mandal et al. - 2023 - Topological superconductors from a materials persp.pdf;/Users/wasmer/Zotero/storage/L8FFWWRW/2303.html}
}
@@ -10912,6 +11940,23 @@ Subject\_term\_id: computational-methods;density-functional-theory;method-develo
file = {/Users/wasmer/Nextcloud/Zotero/Manica et al_2023_Accelerating material design with the generative toolkit for scientific2.pdf}
}
+@incollection{mannStateSpecies2013,
+ title = {State of the {{Species}}},
+ booktitle = {State of the {{Species}}},
+ author = {Mann, Charles C.},
+ date = {2013-12-10},
+ pages = {500--526},
+ publisher = {Columbia University Press},
+ doi = {10.7312/asme16225-019},
+ url = {https://www.degruyter.com/document/doi/10.7312/asme16225-019/html?lang=en},
+ urldate = {2024-08-01},
+ abstract = {State of the Species was published in The Best American Magazine Writing 2013 on page 500.},
+ isbn = {978-0-231-53706-3},
+ langid = {english},
+ keywords = {anthropocene,climate change,energy challenge,essay,for introductions,great acceleration},
+ file = {/Users/wasmer/Nextcloud/Zotero/Mann_2013_State of the Species.pdf}
+}
+
@article{manzoorMachineLearningBased2021,
title = {Machine {{Learning Based Methodology}} to {{Predict Point Defect Energies}} in {{Multi-Principal Element Alloys}}},
author = {Manzoor, Anus and Arora, Gaurav and Jerome, Bryant and Linton, Nathan and Norman, Bailey and Aidhy, Dilpuneet S.},
@@ -10988,7 +12033,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
urldate = {2023-03-20},
abstract = {Machine learning (ML) algorithms are currently emerging as powerful tools in all areas of science. Conventionally, ML is understood as a fundamentally data-driven endeavour. Unfortunately, large well-curated databases are sparse in chemistry. In this contribution, I therefore review science-driven ML approaches which do not rely on “big dataâ€, focusing on the atomistic modelling of materials and molecules. In this context, the term science-driven refers to approaches that begin with a scientific question and then ask what training data and model design choices are appropriate. As key features of science-driven ML, the automated and purpose-driven collection of data and the use of chemical and physical priors to achieve high data-efficiency are discussed. Furthermore, the importance of appropriate model evaluation and error estimation is emphasized.},
langid = {english},
- keywords = {\_tablet,active learning,all-electron,AML,body-order,data-driven,database generation,delta learning,equivariant,inductive bias,iterative learning,iterative learning scheme,MACE,ML,ML-DFA,ML-DFT,ML-ESM,model evaluation,physical prior,physics-informed ML,prediction of electron density,review,review-of-AML,science-driven,uncertainty quantification},
+ keywords = {active learning,all-electron,AML,body-order,data-driven,database generation,delta learning,equivariant,inductive bias,iterative learning,iterative learning scheme,MACE,ML,ML-DFA,ML-DFT,ML-ESM,model evaluation,physical prior,physics-informed ML,prediction of electron density,review,review-of-AML,science-driven,uncertainty quantification},
file = {/Users/wasmer/Zotero/storage/LIPPS6I7/Margraf_2023_Science-Driven Atomistic Machine Learning.pdf;/Users/wasmer/Zotero/storage/V3VTFITJ/ange.html}
}
@@ -11006,7 +12051,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
urldate = {2024-04-07},
abstract = {Machine learning (ML) algorithms are currently emerging as powerful tools in all areas of science. Conventionally, ML is understood as a fundamentally data-driven endeavour. Unfortunately, large well-curated databases are sparse in chemistry. In this contribution, I therefore review science-driven ML approaches which do not rely on “big dataâ€, focusing on the atomistic modelling of materials and molecules. In this context, the term science-driven refers to approaches that begin with a scientific question and then ask what training data and model design choices are appropriate. As key features of science-driven ML, the automated and purpose-driven collection of data and the use of chemical and physical priors to achieve high data-efficiency are discussed. Furthermore, the importance of appropriate model evaluation and error estimation is emphasized.},
langid = {english},
- keywords = {\_tablet,active learning,all-electron,AML,body-order,data-driven,database generation,delta learning,equivariant,inductive bias,iterative learning,iterative learning scheme,MACE,ML,ML-DFA,ML-DFT,ML-ESM,model evaluation,physical prior,physics-informed ML,prediction of electron density,review,review-of-AML,science-driven,uncertainty quantification},
+ keywords = {active learning,all-electron,AML,body-order,data-driven,database generation,delta learning,equivariant,inductive bias,iterative learning,iterative learning scheme,MACE,ML,ML-DFA,ML-DFT,ML-ESM,model evaluation,physical prior,physics-informed ML,prediction of electron density,review,review-of-AML,science-driven,uncertainty quantification},
file = {/Users/wasmer/Nextcloud/Zotero/Margraf_2023_Science-Driven Atomistic Machine Learning.pdf;/Users/wasmer/Zotero/storage/5BDTDNS9/anie.html}
}
@@ -11023,10 +12068,29 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
urldate = {2023-07-04},
abstract = {The study of electronic structure of materials is at a momentous stage, with new computational methods and advances in basic theory. Many properties of materials can be determined from the fundamental equations, and electronic structure theory is now an integral part of research in physics, chemistry, materials science and other fields. This book provides a unified exposition of the theory and methods, with emphasis on understanding each essential component. New in the second edition are recent advances in density functional theory, an introduction to Berry phases and topological insulators explained in terms of elementary band theory, and many new examples of applications. Graduate students and research scientists will find careful explanations with references to original papers, pertinent reviews, and accessible books. Each chapter includes a short list of the most relevant works and exercises that reveal salient points and challenge the reader.},
isbn = {978-1-108-42990-0},
- keywords = {\_tablet},
file = {/Users/wasmer/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Martin_2020_Electronic Structure.pdf;/Users/wasmer/Zotero/storage/PUPKBCZR/ED0FF348536BFFE8899627C8F78FEE6A.html}
}
+@article{martinettoInvertingKohnSham2024,
+ title = {Inverting the {{Kohn}}–{{Sham}} Equations with Physics-Informed Machine Learning},
+ author = {Martinetto, Vincent and Shah, Karan and Cangi, Attila and Pribram-Jones, Aurora},
+ date = {2024-03},
+ journaltitle = {Machine Learning: Science and Technology},
+ shortjournal = {Mach. Learn.: Sci. Technol.},
+ volume = {5},
+ number = {1},
+ pages = {015050},
+ publisher = {IOP Publishing},
+ issn = {2632-2153},
+ doi = {10.1088/2632-2153/ad3159},
+ url = {https://dx.doi.org/10.1088/2632-2153/ad3159},
+ urldate = {2024-06-07},
+ abstract = {Electronic structure theory calculations offer an understanding of matter at the quantum level, complementing experimental studies in materials science and chemistry. One of the most widely used methods, density functional theory, maps a set of real interacting electrons to a set of fictitious non-interacting electrons that share the same probability density. Ensuring that the density remains the same depends on the exchange-correlation (XC) energy and, by a derivative, the XC potential. Inversions provide a method to obtain exact XC potentials from target electronic densities, in hopes of gaining insights into accuracy-boosting approximations. Neural networks provide a new avenue to perform inversions by learning the mapping from density to potential. In this work, we learn this mapping using physics-informed machine learning methods, namely physics informed neural networks and Fourier neural operators. We demonstrate the capabilities of these two methods on a dataset of one-dimensional atomic and molecular models. The capabilities of each approach are discussed in conjunction with this proof-of-concept presentation. The primary finding of our investigation is that the combination of both approaches has the greatest potential for inverting the Kohn–Sham equations at scale.},
+ langid = {english},
+ keywords = {/unread,1D materials,DFT,HK map,inverse DFT,inverse problem,ML,ML-DFA,ML-DFT,ML-ESM,prediction from density,prediction of electron potential,with-code,with-data},
+ file = {/Users/wasmer/Nextcloud/Zotero/Martinetto et al_2024_Inverting the Kohn–Sham equations with physics-informed machine learning.pdf}
+}
+
@article{martinez-carracedoElectricallyDrivenSinglettriplet2023,
title = {Electrically Driven Singlet-Triplet Transition in Triangulene Spin-1 Chains},
author = {MartÃnez-Carracedo, Gabriel and Oroszlány, László and GarcÃa-Fuente, Amador and Szunyogh, László and Ferrer, Jaime},
@@ -11050,13 +12114,13 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {MartÃnez-Carracedo, Gabriel and Oroszlány, László and GarcÃa-Fuente, Amador and Nyári, Bendegúz and Udvardi, László and Szunyogh, László and Ferrer, Jaime},
date = {2023-09-05},
eprint = {2309.02558},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2309.02558},
url = {http://arxiv.org/abs/2309.02558},
urldate = {2023-09-20},
abstract = {We propose a method to determine the magnetic exchange interaction and on-site anisotropy tensors of extended Heisenberg spin models from density functional theory including relativistic effects. The method is based on the Liechtenstein-Katsnelson-Antropov-Gubanov torque formalism, whereby energy variations upon infinitesimal rotations are performed. We assume that the Kohn-Sham Hamiltonian is expanded in a non-orthogonal basis set of pseudo-atomic orbitals. We define local operators that are both hermitian and satisfy relevant sum rules. We demonstrate that in the presence of spin-orbit coupling a correct mapping from the density functional total energy to a spin model that relies on the rotation of the exchange field part of the Hamiltonian can not be accounted for by transforming the full Hamiltonian. We derive a set of sum rules that pose stringent validity tests on any specific calculation. We showcase the flexibility and accuracy of the method by computing the exchange and anisotropy tensors of both well-studied magnetic nanostructures and of recently synthesized two-dimensional magnets. Specifically, we benchmark our approach against the established Korringa-Kohn-Rostoker Green's function method and show that they agree well. Finally, we demonstrate how the application of biaxial strain on the two-dimensional magnet T-CrTe2 can trigger a magnetic phase transition.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Budapest KKR group,DFT,GF2023 workshop,Jij,SIESTA,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/MartÃnez-Carracedo et al_2023_Relativistic magnetic interactions from non-orthogonal basis sets.pdf;/Users/wasmer/Zotero/storage/8KP4SZA4/2309.html}
}
@@ -11075,7 +12139,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
url = {https://link.aps.org/doi/10.1103/PhysRevB.108.214418},
urldate = {2024-01-01},
abstract = {We propose a method to determine the magnetic exchange interaction and onsite anisotropy tensors of extended Heisenberg spin models from density functional theory including relativistic effects. The method is based on the Liechtenstein-Katsnelson-Antropov-Gubanov torque formalism, whereby energy variations upon infinitesimal rotations are performed. We assume that the Kohn-Sham Hamiltonian is expanded in a nonorthogonal basis set of pseudoatomic orbitals. We define local operators that are both Hermitian and satisfy relevant sum rules. We demonstrate that in the presence of spin-orbit coupling a correct mapping from the density functional total energy to a spin model that relies on the rotation of the exchange field part of the Hamiltonian can not be accounted for by transforming the full Hamiltonian. We derive a set of sum rules that pose stringent validity tests on any specific calculation. We showcase the flexibility and accuracy of the method by computing the exchange and anisotropy tensors of both well-studied magnetic nanostructures and of recently synthesized two-dimensional magnets. Specifically, we benchmark our approach against the established Korringa-Kohn-Rostoker Green's function method and show that they agree well. Finally, we demonstrate how the application of biaxial strain on the two-dimensional magnet T−CrTe2 can trigger a magnetic phase transition.},
- keywords = {/unread,Budapest KKR group,DFT,GF2023 workshop,Jij,SIESTA,todo-tagging},
+ keywords = {Budapest KKR group,condensed matter,DFT,exchange interaction,GF2023 workshop,good figures,Heisenberg model,Jij,KKR,learning material,Liechtenstein formula,magnetism,physics,SIESTA,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/MartÃnez-Carracedo et al_2023_Relativistic magnetic interactions from nonorthogonal basis sets.pdf;/Users/wasmer/Zotero/storage/EG37HM8T/PhysRevB.108.html}
}
@@ -11138,18 +12202,36 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Ma, Yuxing and Zhong, Yang and Hongyu, Yu and Chen, Shiyou and Xiang, Hongjun},
date = {2023-06-13},
eprint = {2306.08017},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2306.08017},
url = {http://arxiv.org/abs/2306.08017},
urldate = {2023-10-13},
abstract = {The study of the electronic properties of charged defects is crucial for our understanding of various electrical properties of materials. However, the high computational cost of density functional theory (DFT) hinders the research on large defect models. In this study, we present an E(3) equivariant graph neural network framework (HamGNN-Q), which can predict the tight-binding Hamiltonian matrices for various defect types with different charges using only one set of network parameters. By incorporating features of background charge into the element representation, HamGNN-Q enables a direct mapping from structure and background charge to the electronic Hamiltonian matrix of charged defect systems without DFT calculation. We demonstrate the model's high precision and transferability through testing on GaAs systems with various charged defect configurations. Our approach provides a practical solution for accelerating charged defect electronic structure calculations and advancing the design of materials with tailored electronic properties.},
- pubstate = {preprint},
- keywords = {\_tablet,AML,defects,disordered,HamGNN,ML,ML-DFT,ML-ESM,MPNN,point defects,prediction of Hamiltonian matrix},
+ pubstate = {prepublished},
+ keywords = {AML,defects,disordered,HamGNN,ML,ML-DFT,ML-ESM,MPNN,point defects,prediction of Hamiltonian matrix},
file = {/Users/wasmer/Nextcloud/Zotero/Ma et al_2023_Transferable Machine Learning Approach for Predicting Electronic Structures of.pdf;/Users/wasmer/Zotero/storage/TICQMBV5/2306.html}
}
-@inproceedings{mavropoulosKorringaKohnRostokerKKRGreen2006,
+@article{mavropoulosExchangeCouplingTransitionmetal2010,
+ title = {Exchange Coupling in Transition-Metal Nanoclusters on {{Cu}}(001) and {{Cu}}(111) Surfaces},
+ author = {Mavropoulos, Phivos and Lounis, Samir and Blügel, Stefan},
+ date = {2010},
+ journaltitle = {physica status solidi (b)},
+ volume = {247},
+ number = {5},
+ pages = {1187--1196},
+ issn = {1521-3951},
+ doi = {10.1002/pssb.200945535},
+ url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/pssb.200945535},
+ urldate = {2024-05-24},
+ abstract = {We present results of density-functional calculations on the magnetic properties of Cr, Mn, Fe and Co nanoclusters (1–9 atoms large) supported on Cu(001) and Cu(111). The inter-atomic exchange coupling is found to depend on competing mechanisms, namely ferromagnetic double exchange and antiferromagnetic kinetic exchange. Hybridization-induced broadening of the resonances is shown to be important for the coupling strength. The cluster shape is found to affect the coupling via a mechanism that comprises the different orientation of the atomic d-orbitals and the strength of nearest-neighbour hopping. Especially in Fe clusters, a correlation of binding energy and exchange coupling is also revealed.},
+ langid = {english},
+ keywords = {DFT,exchange interaction,Heisenberg model,infinitesimal rotation,Jij,JuKKR,KKR,Liechtenstein formula,magnetism,nanomaterials,PGI-1/IAS-1,physics,rec-by-ruess,surface physics,transition metals},
+ file = {/Users/wasmer/Nextcloud/Zotero/Mavropoulos et al_2010_Exchange coupling in transition-metal nanoclusters on Cu(001) and Cu(111).pdf;/Users/wasmer/Zotero/storage/IJFWJQ48/pssb.html}
+}
+
+@incollection{mavropoulosKorringaKohnRostokerKKRGreen2006,
title = {The {{Korringa-Kohn-Rostoker}} ({{KKR}}) {{Green}} Function Method {{I}}. {{Electronic}} Structure of Periodic Systems},
booktitle = {Computational {{Nanoscience}}: {{Do It Yourself}}! - {{Lecture Notes}}},
author = {Mavropoulos, Phivos and Papanikolaou, Nikos},
@@ -11157,14 +12239,14 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
series = {{{NIC}} Series},
volume = {31},
pages = {131--158},
- publisher = {NIC-Secretariat, Research Centre Jülich},
+ publisher = {John von Neumann Institute for Computing},
location = {Jülich},
url = {http://hdl.handle.net/2128/4777},
urldate = {2021-06-28},
abstract = {The Korringa-Kohn-Rostoker (KKR) method for the calculation of the electronic structure ofmaterials is founded on the concepts of the Green function and of multiple-scattering. In thismanuscript, after a short introduction to Green functions,we present a description of single-site scattering (including anisotropic potentials) and multiple-scattering theory and the KKRequations. The KKR representation of the Green function andthe algebraic Dyson equation areintroduced. We then discuss the screened KKR formalism, andits advantages in the numericaleffort for the calculation of layered systems. Finally we give a summary of the self-consistencyalgorithm for the calculation of the electronic structure.},
- eventtitle = {Computational {{Nanoscience}}: {{Do It Yourself}}!},
isbn = {3-00-017350-1},
- keywords = {\_tablet,FZJ,KKR,PGI-1/IAS-1},
+ langid = {english},
+ keywords = {\_tablet,all-electron,DFT,DFT theory,educational,Electronic structure,electronic structure theory,full-potential,full-relativistic,FZJ,JuKKR,KKR,KKR foundations,Korringa–Kohn–Rostoker,learning material,library,PGI-1/IAS-1,with-code},
annotation = {Johannes Grotendorst, Stefan Blügel, Dominik Marx (Editors)},
file = {/Users/wasmer/Nextcloud/Zotero/Mavropoulos_Papanikolaou_2006_The Korringa-Kohn-Rostoker (KKR) Green function method I.pdf}
}
@@ -11205,7 +12287,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
url = {https://www.nature.com/articles/s43588-023-00562-8},
urldate = {2024-03-08},
langid = {english},
- keywords = {\_tablet,AML,density matrix,editorial highlight,ML,ML-DFT,ML-ESM,ML-WFT,prediction of density matrix,prediction of electron density},
+ keywords = {AML,density matrix,editorial highlight,ML,ML-DFT,ML-ESM,ML-WFT,prediction of density matrix,prediction of electron density},
file = {/Users/wasmer/Nextcloud/Zotero/McCardle_2023_Predicting electronic structure calculation results.pdf}
}
@@ -11260,7 +12342,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
abstract = {A key challenge in the effort to simulate today’s quantum computing devices is the ability to learn and encode the complex correlations that occur between qubits. Emerging technologies based on language models adopted from machine learning have shown unique abilities to learn quantum states. We highlight the contributions that language models are making in the effort to build quantum computers and discuss their future role in the race to quantum advantage.},
issue = {1},
langid = {english},
- keywords = {/unread,GPT,language models,LLM,Quantum simulation,RNN,todo-tagging},
+ keywords = {GPT,language models,LLM,Quantum simulation,RNN,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Melko_Carrasquilla_2024_Language models for quantum simulation.pdf}
}
@@ -11290,13 +12372,13 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Mellor, Joseph and Turner, Jack and Storkey, Amos and Crowley, Elliot J.},
date = {2021-06-11},
eprint = {2006.04647},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.2006.04647},
url = {http://arxiv.org/abs/2006.04647},
urldate = {2023-10-05},
abstract = {The time and effort involved in hand-designing deep neural networks is immense. This has prompted the development of Neural Architecture Search (NAS) techniques to automate this design. However, NAS algorithms tend to be slow and expensive; they need to train vast numbers of candidate networks to inform the search process. This could be alleviated if we could partially predict a network's trained accuracy from its initial state. In this work, we examine the overlap of activations between datapoints in untrained networks and motivate how this can give a measure which is usefully indicative of a network's trained performance. We incorporate this measure into a simple algorithm that allows us to search for powerful networks without any training in a matter of seconds on a single GPU, and verify its effectiveness on NAS-Bench-101, NAS-Bench-201, NATS-Bench, and Network Design Spaces. Our approach can be readily combined with more expensive search methods; we examine a simple adaptation of regularised evolutionary search. Code for reproducing our experiments is available at https://github.com/BayesWatch/nas-without-training.},
- pubstate = {preprint},
+ pubstate = {prepublished},
version = {3},
keywords = {autoML,Deep learning,General ML,hyperparameters,hyperparameters optimization,library,MALA,ML,NN,NN architecture,PyTorch,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Mellor et al_2021_Neural Architecture Search without Training.pdf;/Users/wasmer/Zotero/storage/NKM9IWFY/2006.html}
@@ -11372,7 +12454,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
urldate = {2022-10-17},
abstract = {The determination of magnetic structure poses a long-standing challenge in condensed matter physics and materials science. Experimental techniques such as neutron diffraction are resource-limited and require complex structure refinement protocols, while computational approaches such as first-principles density functional theory (DFT) need additional semi-empirical correction, and reliable prediction is still largely limited to collinear magnetism. Here, we present a machine learning model that aims to classify the magnetic structure by inputting atomic coordinates containing transition metal and rare earth elements. By building a Euclidean equivariant neural network that preserves the crystallographic symmetry, the magnetic structure (ferromagnetic, antiferromagnetic, and non-magnetic) and magnetic propagation vector (zero or non-zero) can be predicted with an average accuracy of 77.8\% and 73.6\%. In particular, a 91\% accuracy is reached when predicting no magnetic ordering even if the structure contains magnetic element(s). Our work represents one step forward to solving the grand challenge of full magnetic structure determination.},
langid = {english},
- keywords = {\_tablet,AFM,AML,classification,classification of magnetic structure,collinear,e3nn,electronegativity,equivariant,FM,GNN,library,MAGNDATA,magnetic moment,magnetic order,magnetic structure,magnetism,magnetism database,materials,materials project,ML,MPNN,non-collinear,polarizability,prediction of magnetic order,propagation vector,rare earths,spin-dependent,transition metals,vectorial learning target,with-code},
+ keywords = {AFM,AML,classification,classification of magnetic structure,collinear,e3nn,electronegativity,equivariant,FM,GNN,library,MAGNDATA,magnetic moment,magnetic order,magnetic structure,magnetism,magnetism database,materials,materials project,ML,MPNN,non-collinear,polarizability,prediction of magnetic order,propagation vector,rare earths,spin-dependent,transition metals,vectorial learning target,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Merker et al_2022_Machine learning magnetism classifiers from atomic coordinates.pdf;/Users/wasmer/Zotero/storage/7UQX89UL/S258900422201464X.html}
}
@@ -11401,13 +12483,13 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Metz, Luke and Harrison, James and Freeman, C. Daniel and Merchant, Amil and Beyer, Lucas and Bradbury, James and Agrawal, Naman and Poole, Ben and Mordatch, Igor and Roberts, Adam and Sohl-Dickstein, Jascha},
date = {2022-11-17},
eprint = {2211.09760},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math, stat},
doi = {10.48550/arXiv.2211.09760},
url = {http://arxiv.org/abs/2211.09760},
urldate = {2023-07-21},
abstract = {While deep learning models have replaced hand-designed features across many domains, these models are still trained with hand-designed optimizers. In this work, we leverage the same scaling approach behind the success of deep learning to learn versatile optimizers. We train an optimizer for deep learning which is itself a small neural network that ingests gradients and outputs parameter updates. Meta-trained with approximately four thousand TPU-months of compute on a wide variety of optimization tasks, our optimizer not only exhibits compelling performance, but optimizes in interesting and unexpected ways. It requires no hyperparameter tuning, instead automatically adapting to the specifics of the problem being optimized. We open source our learned optimizer, meta-training code, the associated train and test data, and an extensive optimizer benchmark suite with baselines at velo-code.github.io.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {autoML,general ML,Google,hyperparameters,hyperparameters optimization,meta-training,ML,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Metz et al_2022_VeLO.pdf;/Users/wasmer/Zotero/storage/82NVCST9/2211.html}
}
@@ -11435,17 +12517,50 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Minotakis, Michael and Rossignol, Hugo and Cobelli, Matteo and Sanvito, Stefano},
date = {2023-03-29},
eprint = {2303.16597},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2303.16597},
url = {http://arxiv.org/abs/2303.16597},
urldate = {2023-04-13},
abstract = {The prediction of phase diagrams in the search for new phases is a complex and computationally intensive task. Density functional theory provides, in many situations, the desired accuracy, but its throughput becomes prohibitively limited as the number of species involved grows, even when used with local and semi-local functionals. Here, we explore the possibility of integrating machine-learning models in the workflow for the construction of ternary convex hull diagrams. In particular, we train a set of spectral neighbour-analysis potentials (SNAPs) over readily available binary phases and we establish whether this is good enough to predict the energies of novel ternaries. Such a strategy does not require any new calculations specific for the construction of the model, but just avails of data stored in binary-phase-diagram repositories. We find that a so-constructed SNAP is capable of accurate total-energy estimates for ternary phases close to the equilibrium geometry but, in general, is not able to perform atomic relaxation. This is because during a typical relaxation path a given phase traverses regions in the parameter space poorly represented by the training set. Different metrics are then investigated to assess how an unknown structure is well described by a given SNAP model, and we find that the standard deviation of an ensemble of SNAPs provides a fast and non-specie-specific metric.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {\_tablet,AFLOW,AFLOWLIB,AML,binary systems,convex hull,High-throughput,LAMMPS,materials discovery,materials screening,ML,MLP,MTP,PCA,phase diagram,prediction of energy,scikit-learn,SNAP,structure relaxation,ternary systems,VASP},
file = {/Users/wasmer/Nextcloud/Zotero/Minotakis et al_2023_Machine-Learning Surrogate Model for Accelerating the Search of Stable Ternary.pdf;/Users/wasmer/Zotero/storage/L8VGVV3E/2303.html}
}
+@article{mirhosseiniAutomatedApproachDeveloping2021,
+ title = {An Automated Approach for Developing Neural Network Interatomic Potentials with {{FLAME}}},
+ author = {Mirhosseini, Hossein and Tahmasbi, Hossein and Kuchana, Sai Ram and Ghasemi, S. Alireza and Kühne, Thomas D.},
+ date = {2021-09-01},
+ journaltitle = {Computational Materials Science},
+ shortjournal = {Computational Materials Science},
+ volume = {197},
+ pages = {110567},
+ issn = {0927-0256},
+ doi = {10.1016/j.commatsci.2021.110567},
+ url = {https://www.sciencedirect.com/science/article/pii/S0927025621002949},
+ urldate = {2024-06-07},
+ abstract = {The performance of machine learning interatomic potentials relies on the quality of the training dataset. In this work, we present an approach for generating diverse and representative training data points which initiates with ab initio calculations for bulk structures. The data generation and potential construction further proceed side-by-side in a cyclic process of training the neural network and crystal structure prediction based on the developed interatomic potentials. All steps of the data generation and potential development are performed with minimal human intervention. We show the reliability of our approach by assessing the performance of neural network potentials developed for two inorganic systems.},
+ keywords = {ACSF,active learning,AML,CASUS,FLAME,HDNNP,materials project,minima hopping,ML,MLP,PES,prediction of EOS,prediction of formation energy,prediction of phonon dispersion,structure prediction,structure relaxation,with-code,workflows},
+ file = {/Users/wasmer/Nextcloud/Zotero/Mirhosseini et al_2021_An automated approach for developing neural network interatomic potentials with.pdf;/Users/wasmer/Zotero/storage/V88VNAER/S0927025621002949.html}
+}
+
+@article{mirkinEnergyTransitionNeeds2024,
+ title = {Energy Transition Needs New Materials},
+ author = {Mirkin, Chad A. and Sargent, Edward H. and Schrag, Daniel P.},
+ date = {2024-05-17},
+ journaltitle = {Science},
+ volume = {384},
+ number = {6697},
+ pages = {713--713},
+ publisher = {American Association for the Advancement of Science},
+ doi = {10.1126/science.adq3799},
+ url = {https://www.science.org/doi/10.1126/science.adq3799},
+ urldate = {2024-05-25},
+ keywords = {AI4Science,energy consumption,energy efficiency,for introductions,materials,opinion},
+ file = {/Users/wasmer/Nextcloud/Zotero/Mirkin et al_2024_Energy transition needs new materials.pdf}
+}
+
@inproceedings{missierW3CPROVFamily2013,
title = {The {{W3C PROV}} Family of Specifications for Modelling Provenance Metadata},
booktitle = {Proceedings of the 16th {{International Conference}} on {{Extending Database Technology}}},
@@ -11509,10 +12624,22 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
url = {https://doi.org/10.1021/acs.jpclett.2c03670},
urldate = {2023-04-04},
abstract = {We present an analysis of the static exchange-correlation (XC) kernel computed from hybrid functionals with a single mixing coefficient such as PBE0 and PBE0–1/3. We break down the hybrid XC kernels into the exchange and correlation parts using the Hartree–Fock functional, the exchange-only PBE, and the correlation-only PBE. This decomposition is combined with exact data for the static XC kernel of the uniform electron gas and an Airy gas model within a subsystem functional approach. This gives us a tool for the non-empirical choice of the mixing coefficient under ambient and extreme conditions. Our analysis provides physical insights into the effect of the variation of the mixing coefficient in hybrid functionals, which is of immense practical value. The presented approach is general and can be used for other types of functionals like screened hybrids.},
- keywords = {/unread,\_tablet,CASUS,DFA,DFT,HZDR,PGI-1/IAS-1},
+ keywords = {/unread,CASUS,DFA,DFT,HZDR,PGI-1/IAS-1},
file = {/Users/wasmer/Nextcloud/Zotero/Moldabekov et al_2023_Non-empirical Mixing Coefficient for Hybrid XC Functionals from Analysis of the.pdf;/Users/wasmer/Zotero/storage/WGXJ5PMF/acs.jpclett.html}
}
+@unpublished{molenkampUserGuideSpherical2016,
+ title = {A {{User}}’s {{Guide}} to {{Spherical Harmonics}}},
+ author = {Molenkamp, Martin J.},
+ date = {2016-10-18},
+ location = {Ohio State University},
+ url = {http://www.ohiouniversityfaculty.com/mohlenka/research/uguide.pdf},
+ howpublished = {Pamphlet},
+ pagetotal = {14},
+ keywords = {educational,Hilbert space,learning material,mathematics,numerical analysis,numerical methods,spherical harmonics,theory,tutorial},
+ file = {/Users/wasmer/Nextcloud/Zotero/Molenkamp_2016_A User’s Guide to Spherical Harmonics.pdf}
+}
+
@book{molnarGlobalSurrogateInterpretable,
title = {5.6 {{Global Surrogate}} | {{Interpretable Machine Learning}}},
author = {Molnar, Christoph},
@@ -11651,14 +12778,14 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Morrow, Joe D. and Gardner, John L. A. and Deringer, Volker L.},
date = {2022-11-28},
eprint = {2211.12484},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2211.12484},
url = {http://arxiv.org/abs/2211.12484},
urldate = {2023-01-02},
abstract = {Machine learning (ML) approaches enable large-scale atomistic simulations with near-quantum-mechanical accuracy. With the growing availability of these methods there arises a need for careful validation, particularly for physically agnostic models - that is, for potentials which extract the nature of atomic interactions from reference data. Here, we review the basic principles behind ML potentials and their validation for atomic-scale materials modeling. We discuss best practice in defining error metrics based on numerical performance as well as physically guided validation. We give specific recommendations that we hope will be useful for the wider community, including those researchers who intend to use ML potentials for materials "off the shelf".},
- pubstate = {preprint},
- keywords = {\_tablet,benchmarking,best practices,cross-validation,experimental data,GAP,how-to,ML,MLP,model evaluation,model validation,models comparison,MTP,numerical errors,SOAP,SOTA,tutorial},
+ pubstate = {prepublished},
+ keywords = {benchmarking,best practices,cross-validation,experimental data,GAP,how-to,ML,MLP,model evaluation,model validation,models comparison,MTP,numerical errors,SOAP,SOTA,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Morrow et al_2022_How to validate machine-learned interatomic potentials.pdf;/Users/wasmer/Zotero/storage/TW3TCHB3/2211.html}
}
@@ -11727,7 +12854,22 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
langid = {english},
pagetotal = {85},
keywords = {\_tablet,AiiDA,aiida-kkr,Hall QAHE,impurity embedding,juKKR,KKR,master-thesis,PGI-1/IAS-1,thesis},
- file = {/Users/wasmer/Nextcloud/Zotero/Mozumder_2022_Design of magnetic interactions in doped topological insulators.pdf}
+ file = {/Users/wasmer/Nextcloud/Zotero/Mozumder et al_2022_Design of magnetic interactions in doped topological insulators.pdf}
+}
+
+@online{mozumderHighthroughputMagneticCodoping2024,
+ title = {High-Throughput Magnetic Co-Doping and Design of Exchange Interactions in a Topological Insulator},
+ author = {Mozumder, Rubel and Wasmer, Johannes and Silva, David Antognini and Blügel, Stefan and Rüßmann, Philipp},
+ date = {2024-07-05},
+ eprint = {2407.04413},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat},
+ url = {http://arxiv.org/abs/2407.04413},
+ urldate = {2024-07-08},
+ abstract = {Using high-throughput automation of ab-initio impurity embedding simulations, we created a database of \$3d\$ and \$4d\$ transition metal defects embedded into the prototypical topological insulator (TI) Bi\$\_2\$Te\$\_3\$. We simulate both single impurities as well as impurity dimers at different impurity-impurity distances inside the TI. We extract changes to magnetic moments, analyze the polarizability of non-magnetic impurity atoms via nearby magnetic impurity atoms and calculate the exchange coupling constants for a Heisenberg Hamiltonian. We uncover chemical trends in the exchange coupling constants and discuss the impurities' potential with respect to magnetic order in the fields of quantum anomalous Hall insulators and topological quantum computing. In particular, we predict that co-doping of different magnetic dopants is a viable strategy to engineer the magnetic ground state in magnetic TIs.},
+ pubstate = {prepublished},
+ keywords = {AiiDA,aiida-kkr,Bi2Te3,co-doping,database generation,dataset,defects,exchange interaction,FZJ,Heisenberg model,impurity embedding,Jij,JuDFT,juKKR,KKR,Liechtenstein formula,magnetic topological materials,magnetism,Materials Cloud,PGI,PGI-1/IAS-1,physics,point defects,topological,topological insulator,with-data},
+ file = {/Users/wasmer/Nextcloud/Zotero/Mozumder et al_2024_High-throughput magnetic co-doping and design of exchange interactions in a.pdf;/Users/wasmer/Zotero/storage/GXLBFXQF/2407.html}
}
@online{muckleyInterpretableModelsExtrapolation2022,
@@ -11735,13 +12877,13 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Muckley, Eric S. and Saal, James E. and Meredig, Bryce and Roper, Christopher S. and Martin, John H.},
date = {2022-12-16},
eprint = {2212.10283},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2212.10283},
url = {http://arxiv.org/abs/2212.10283},
urldate = {2023-08-19},
abstract = {Data-driven models are central to scientific discovery. In efforts to achieve state-of-the-art model accuracy, researchers are employing increasingly complex machine learning algorithms that often outperform simple regressions in interpolative settings (e.g. random k-fold cross-validation) but suffer from poor extrapolation performance, portability, and human interpretability, which limits their potential for facilitating novel scientific insight. Here we examine the trade-off between model performance and interpretability across a broad range of science and engineering problems with an emphasis on materials science datasets. We compare the performance of black box random forest and neural network machine learning algorithms to that of single-feature linear regressions which are fitted using interpretable input features discovered by a simple random search algorithm. For interpolation problems, the average prediction errors of linear regressions were twice as high as those of black box models. Remarkably, when prediction tasks required extrapolation, linear models yielded average error only 5\% higher than that of black box models, and outperformed black box models in roughly 40\% of the tested prediction tasks, which suggests that they may be desirable over complex algorithms in many extrapolation problems because of their superior interpretability, computational overhead, and ease of use. The results challenge the common assumption that extrapolative models for scientific machine learning are constrained by an inherent trade-off between performance and interpretability.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Citrine Informatics,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Muckley et al_2022_Interpretable models for extrapolation in scientific machine learning.pdf;/Users/wasmer/Zotero/storage/HE59WRRZ/2212.html}
}
@@ -11751,13 +12893,13 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Müller, Luis and Galkin, Mikhail and Morris, Christopher and Rampášek, Ladislav},
date = {2023-02-08},
eprint = {2302.04181},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2302.04181},
url = {http://arxiv.org/abs/2302.04181},
urldate = {2023-07-24},
abstract = {Recently, transformer architectures for graphs emerged as an alternative to established techniques for machine learning with graphs, such as graph neural networks. So far, they have shown promising empirical results, e.g., on molecular prediction datasets, often attributed to their ability to circumvent graph neural networks' shortcomings, such as over-smoothing and over-squashing. Here, we derive a taxonomy of graph transformer architectures, bringing some order to this emerging field. We overview their theoretical properties, survey structural and positional encodings, and discuss extensions for important graph classes, e.g., 3D molecular graphs. Empirically, we probe how well graph transformers can recover various graph properties, how well they can deal with heterophilic graphs, and to what extent they prevent over-squashing. Further, we outline open challenges and research direction to stimulate future work. Our code is available at https://github.com/luis-mueller/probing-graph-transformers.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,attention,benchmarking,general ML,GNN,graph attention,graph ML,graph transformer,ML,OGB,transformer},
file = {/Users/wasmer/Nextcloud/Zotero/Müller et al_2023_Attending to Graph Transformers.pdf;/Users/wasmer/Zotero/storage/I3F8C5FS/2302.html}
}
@@ -11777,7 +12919,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
url = {https://link.aps.org/doi/10.1103/PhysRevB.99.224414},
urldate = {2023-10-26},
abstract = {The Spirit framework is designed for atomic-scale spin simulations of magnetic systems with arbitrary geometry and magnetic structure, providing a graphical user interface with powerful visualizations and an easy-to-use scripting interface. An extended Heisenberg-type spin-lattice Hamiltonian including competing exchange interactions between neighbors at arbitrary distances, higher-order exchange, Dzyaloshinskii-Moriya and dipole-dipole interactions is used to describe the energetics of a system of classical spins localized at atom positions. A variety of common simulation methods are implemented including Monte Carlo and various time evolution algorithms based on the Landau-Lifshitz-Gilbert (LLG) equation of motion. These methods can be used to determine static ground-state and metastable spin configurations, sample equilibrium and finite-temperature thermodynamical properties of magnetic materials and nanostructures, or calculate dynamical trajectories including spin torques induced by stochastic temperature or electric current. Methods for finding the mechanism and rate of thermally assisted transitions include the geodesic nudged elastic band method, which can be applied when both initial and final states are specified, and the minimum mode-following method when only the initial state is given. The lifetimes of magnetic states and rates of transitions can be evaluated within the harmonic approximation of transition-state theory. The framework offers performant central processing unit (CPU) and graphics processing unit (GPU) parallelizations. All methods are verified and applications to several systems, such as vortices, domain walls, skyrmions, and bobbers are described.},
- keywords = {/unread,FZJ,Heisenberg model,Jij,Landau-Lifshits-Gilbert equation,library,magnetic moment,magnetism,PGI,PGI-1/IAS-1,physics,spin dynamics,spin texture,spin-dependent,Spirit,with-code},
+ keywords = {FZJ,Heisenberg model,Jij,Landau-Lifshits-Gilbert equation,library,magnetic moment,magnetism,PGI,PGI-1/IAS-1,physics,spin dynamics,spin texture,spin-dependent,Spirit,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Müller et al_2019_Spirit.pdf;/Users/wasmer/Zotero/storage/N8F8Z9GD/PhysRevB.99.html}
}
@@ -11795,7 +12937,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Musaelian, Albert and Batzner, Simon and Johansson, Anders and Sun, Lixin and Owen, Cameron J. and Kornbluth, Mordechai and Kozinsky, Boris},
date = {2022-04-11},
eprint = {2204.05249},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2204.05249},
urldate = {2022-04-14},
@@ -11844,13 +12986,13 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
author = {Musaelian, Albert and Johansson, Anders and Batzner, Simon and Kozinsky, Boris},
date = {2023-04-19},
eprint = {2304.10061},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, q-bio},
doi = {10.48550/arXiv.2304.10061},
url = {http://arxiv.org/abs/2304.10061},
urldate = {2023-09-04},
abstract = {This work brings the leading accuracy, sample efficiency, and robustness of deep equivariant neural networks to the extreme computational scale. This is achieved through a combination of innovative model architecture, massive parallelization, and models and implementations optimized for efficient GPU utilization. The resulting Allegro architecture bridges the accuracy-speed tradeoff of atomistic simulations and enables description of dynamics in structures of unprecedented complexity at quantum fidelity. To illustrate the scalability of Allegro, we perform nanoseconds-long stable simulations of protein dynamics and scale up to a 44-million atom structure of a complete, all-atom, explicitly solvated HIV capsid on the Perlmutter supercomputer. We demonstrate excellent strong scaling up to 100 million atoms and 70\% weak scaling to 5120 A100 GPUs.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Allegro,AML,biomolecules,Gordon Bell Prize,HPC,large-scale simulation,ML,MLP,scaling},
file = {/Users/wasmer/Nextcloud/Zotero/Musaelian et al_2023_Scaling the leading accuracy of deep equivariant models to biomolecular.pdf;/Users/wasmer/Zotero/storage/QQDA94V4/2304.html}
}
@@ -11924,21 +13066,21 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
url = {https://doi.org/10.1021/acs.chemrev.1c00021},
urldate = {2021-08-16},
abstract = {The first step in the construction of a regression model or a data-driven analysis, aiming to predict or elucidate the relationship between the atomic-scale structure of matter and its properties, involves transforming the Cartesian coordinates of the atoms into a suitable representation. The development of atomic-scale representations has played, and continues to play, a central role in the success of machine-learning methods for chemistry and materials science. This review summarizes the current understanding of the nature and characteristics of the most commonly used structural and chemical descriptions of atomistic structures, highlighting the deep underlying connections between different frameworks and the ideas that lead to computationally efficient and universally applicable models. It emphasizes the link between properties, structures, their physical chemistry, and their mathematical description, provides examples of recent applications to a diverse set of chemical and materials science problems, and outlines the open questions and the most promising research directions in the field.},
- keywords = {\_tablet,ACSF,descriptor comparison,descriptor dimred,descriptors,MBTR,ML,review,SOAP},
+ keywords = {ACSF,descriptor comparison,descriptor dimred,descriptors,MBTR,ML,review,SOAP},
file = {/Users/wasmer/Nextcloud/Zotero/Musil et al_2021_Physics-Inspired Structural Representations for Molecules and Materials.pdf}
}
-@unpublished{musilPhysicsinspiredStructuralRepresentations2021,
+@unpublished{musilPhysicsinspiredStructuralRepresentations2021a,
title = {Physics-Inspired Structural Representations for Molecules and Materials},
author = {Musil, Felix and Grisafi, Andrea and Bartók, Albert P. and Ortner, Christoph and Csányi, Gábor and Ceriotti, Michele},
date = {2021-05-04},
eprint = {2101.04673},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
url = {http://arxiv.org/abs/2101.04673},
urldate = {2021-05-30},
abstract = {The first step in the construction of a regression model or a data-driven analysis framework for matter at the atomic scale involves transforming the Cartesian coordinates that describe the positions of the atoms in the form of a representation that obeys the same symmetries as the properties of interest, and in general reflects the physical nature of the problem. The link between properties, structures, their physical chemistry and their mathematical description is strongest when it comes to applications aimed at determining a precise correspondence between atomic configurations and the quantities that one might compute by a quantum mechanical electronic-structure calculation or measure experimentally. The development of atomic-scale representations have played, and continue to play, a central role in the success of machine-learning methods that rely on this correspondence, such as interatomic potentials, as well as generic property models, structural classifiers and low-dimensional maps of structures and datasets. This review summarizes the current understanding of the nature and characteristics of the most commonly used structural and chemical descriptions of molecules and materials, highlighting the deep underlying connections between different frameworks, and the ideas that lead to computationally efficient and universally applicable models. It gives examples of recent applications to a diverse set of chemical and materials science problems, and outlines the open questions and the most promising research directions in the field.},
- keywords = {\_tablet,ACSF,descriptor dimred,descriptors,descriptors analysis,MBTR,ML,review,SOAP},
+ keywords = {ACSF,descriptor dimred,descriptors,descriptors analysis,MBTR,ML,review,SOAP},
file = {/Users/wasmer/Nextcloud/Zotero/Musil et al_2021_Physics-inspired structural representations for molecules and materials.pdf;/Users/wasmer/Zotero/storage/EXTUHGNH/2101.html}
}
@@ -11970,7 +13112,7 @@ Subject\_term\_id: computational-chemistry;density-functional-theory;method-deve
urldate = {2023-10-06},
abstract = {Functional materials that enable many technological applications in our everyday lives owe their unique properties to defects that are carefully engineered and incorporated into these materials during processing. However, optimizing and characterizing these defects is very challenging in practice, making computational modelling an indispensable complementary tool. We have developed an automated workflow and code to accelerate these calculations (AiiDA-defects), which utilises the AiiDA framework, a robust open-source high-throughput materials informatics infrastructure that provides workflow automation while simultaneously preserving and storing the full data provenance in a relational database that is queryable and traversable. This paper describes the design and implementation details of AiiDA-defects, the models and algorithms used, and demonstrates its use in an application to fully characterize the defect chemistry of the well known solid-state Li-ion conductors LiZnPS4. We anticipate that AiiDA-defects will be useful as a tool for fully automated and reproducible defect calculations, allowing detailed defect chemistry to be obtained in a reliable and high-throughput way, and paving the way toward the generation of defects databases for accelerated materials design and discovery.},
langid = {english},
- keywords = {\_tablet,AiiDA,defect chemistry,defects,disordered,FAIR,library,materials informatics,PBC,periodic,point defects,Quantum ESPRESSO,supercell,with-code,workflows},
+ keywords = {AiiDA,database generation,defect chemistry,defects,disordered,FAIR,HTC,library,materials informatics,PBC,periodic,point defects,Quantum ESPRESSO,supercell,with-code,workflows},
file = {/Users/wasmer/Nextcloud/Zotero/Muy et al_2023_AiiDA-defects.pdf}
}
@@ -12106,13 +13248,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Neklyudov, Kirill and Nys, Jannes and Thiede, Luca and Carrasquilla, Juan and Liu, Qiang and Welling, Max and Makhzani, Alireza},
date = {2023-07-16},
eprint = {2307.07050},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2307.07050},
url = {http://arxiv.org/abs/2307.07050},
urldate = {2023-08-22},
abstract = {Solving the quantum many-body Schr\textbackslash "odinger equation is a fundamental and challenging problem in the fields of quantum physics, quantum chemistry, and material sciences. One of the common computational approaches to this problem is Quantum Variational Monte Carlo (QVMC), in which ground-state solutions are obtained by minimizing the energy of the system within a restricted family of parameterized wave functions. Deep learning methods partially address the limitations of traditional QVMC by representing a rich family of wave functions in terms of neural networks. However, the optimization objective in QVMC remains notoriously hard to minimize and requires second-order optimization methods such as natural gradient. In this paper, we first reformulate energy functional minimization in the space of Born distributions corresponding to particle-permutation (anti-)symmetric wave functions, rather than the space of wave functions. We then interpret QVMC as the Fisher-Rao gradient flow in this distributional space, followed by a projection step onto the variational manifold. This perspective provides us with a principled framework to derive new QMC algorithms, by endowing the distributional space with better metrics, and following the projected gradient flow induced by those metrics. More specifically, we propose "Wasserstein Quantum Monte Carlo" (WQMC), which uses the gradient flow induced by the Wasserstein metric, rather than Fisher-Rao metric, and corresponds to transporting the probability mass, rather than teleporting it. We demonstrate empirically that the dynamics of WQMC results in faster convergence to the ground state of molecular systems.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Microsoft Research,ML,ML-QM,ML-QMBP,ML-WFT,prediction of wavefunction,QMC,todo-tagging,VMC},
file = {/Users/wasmer/Nextcloud/Zotero/Neklyudov et al_2023_Wasserstein Quantum Monte Carlo.pdf;/Users/wasmer/Zotero/storage/5BUA924B/2307.html}
}
@@ -12126,7 +13268,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
volume = {106},
number = {4},
eprint = {2201.11647},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:quant-ph},
pages = {045402},
issn = {2469-9950, 2469-9969},
@@ -12152,7 +13294,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
url = {https://link.aps.org/doi/10.1103/PhysRevB.99.075132},
urldate = {2023-03-09},
abstract = {The solution of complex many-body lattice models can often be found by defining an energy functional of the relevant density of the problem. For instance, in the case of the Hubbard model the spin-resolved site occupation is enough to describe the system's total energy. Similarly to standard density functional theory, however, the exact functional is unknown, and suitable approximations need to be formulated. By using a deep-learning neural network trained on exact-diagonalization results, we demonstrate that one can construct an exact functional for the Hubbard model. In particular, we show that the neural network returns a ground-state energy numerically indistinguishable from that obtained by exact diagonalization and, most importantly, that the functional satisfies the two Hohenberg-Kohn theorems: for a given ground-state density it yields the external potential, and it is fully variational in the site occupation.},
- keywords = {/unread,exact diagonaiization,Hubbard model,ML,ML-DFA,ML-DFT,ML-ESM,NN,prediction from density,prediction of ground-state properties,prediction of total energy,spin-dependent},
+ keywords = {/unread,\_tablet,exact diagonaiization,Hubbard model,ML,ML-DFA,ML-DFT,ML-ESM,NN,prediction from density,prediction of ground-state properties,prediction of total energy,spin-dependent},
file = {/Users/wasmer/Nextcloud/Zotero/Nelson et al_2019_Machine learning density functional theory for the Hubbard model.pdf;/Users/wasmer/Zotero/storage/NRLX8AEL/PhysRevB.99.html}
}
@@ -12165,7 +13307,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
volume = {103},
number = {24},
eprint = {2103.05510},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:quant-ph},
pages = {245111},
issn = {2469-9950, 2469-9969},
@@ -12186,7 +13328,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
volume = {3},
number = {10},
eprint = {1906.08534},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
pages = {104405},
issn = {2475-9953},
@@ -12287,13 +13429,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Nigam, Jigyasa and Pozdnyakov, Sergey N. and Huguenin-Dumittan, Kevin K. and Ceriotti, Michele},
date = {2023-02-28},
eprint = {2302.14770},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2302.14770},
url = {http://arxiv.org/abs/2302.14770},
urldate = {2023-03-01},
abstract = {Achieving a complete and symmetric description of a group of point particles, such as atoms in a molecule, is a common problem in physics and theoretical chemistry. The introduction of machine learning to science has made this issue even more critical, as it underpins the ability of a model to reproduce arbitrary physical relationships, and to do so while being consistent with basic symmetries and conservation laws. However, the descriptors that are commonly used to represent point clouds -- most notably those adopted to describe matter at the atomic scale -- are unable to distinguish between special arrangements of particles. This makes it impossible to machine learn their properties. Frameworks that are provably complete exist, but are only so in the limit in which they simultaneously describe the mutual relationship between all atoms, which is impractical. We introduce, and demonstrate on a particularly insidious class of atomic arrangements, a strategy to build descriptors that rely solely on information on the relative arrangement of triplets of particles, but can be used to construct symmetry-adapted models that have universal approximation power.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,ACDC,descriptors,GNN,incompleteness,representation learning},
file = {/Users/wasmer/Nextcloud/Zotero/Nigam et al_2023_Completeness of Atomic Structure Representations.pdf;/Users/wasmer/Zotero/storage/T3WEIUD6/2302.html}
}
@@ -12322,12 +13464,12 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Nigam, Jigyasa and Fraux, Guillaume and Ceriotti, Michele},
date = {2022-02-03},
eprint = {2202.01566},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
url = {http://arxiv.org/abs/2202.01566},
urldate = {2022-02-04},
abstract = {Data-driven schemes that associate molecular and crystal structures with their microscopic properties share the need for a concise, effective description of the arrangement of their atomic constituents. Many types of models rely on descriptions of atom-centered environments, that are associated with an atomic property or with an atomic contribution to an extensive macroscopic quantity. Frameworks in this class can be understood in terms of atom-centered density correlations (ACDC), that are used as a basis for a body-ordered, symmetry-adapted expansion of the targets. Several other schemes, that gather information on the relationship between neighboring atoms using graph-convolutional (or message-passing) ideas, cannot be directly mapped to correlations centered around a single atom. We generalize the ACDC framework to include multi-centered information, generating representations that provide a complete linear basis to regress symmetric functions of atomic coordinates, and form the basis to systematize our understanding of both atom-centered and graph-convolutional machine-learning schemes.},
- keywords = {\_tablet,ACDC,ACE,descriptors,GCN,GNN,ML,MPNN,NequIP,NN,representation learning,SOAP,unified theory},
+ keywords = {ACDC,ACE,descriptors,GCN,GNN,ML,MPNN,NequIP,NN,representation learning,SOAP,unified theory},
file = {/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Nigam et al_2022_Unified theory of atom-centered representations and graph convolutional.pdf}
}
@@ -12346,7 +13488,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
url = {https://link.aps.org/doi/10.1103/PhysRevB.106.235114},
urldate = {2023-04-04},
abstract = {An efficient and accurate generalization of the removed-sphere method (RSM) to solve the Poisson equation for total charge density in a solid with space-filling convex Voronoi polyhedra (VPs) and any symmetry is presented. The generalized RSM avoids the use of multipoles and VP shape functions for cellular integrals, which have associated ill-convergent large, double-internal L sums in spherical-harmonic expansions, so that fast convergence in single-L sums is reached. Our RSM adopts full Ewald formulation to work for all configurations or when symmetry breaking occurs, such as for atomic displacements or elastic constant calculations. The structure-dependent coefficients AL that define RSM can be calculated once for a fixed structure and speed up the whole self-consistent-field procedure. The accuracy and rapid convergence properties are confirmed using two analytic models, including the Coulomb potential and energy. We then implement the full-potential RSM using the Green's function Korringa-Kohn-Rostoker (KKR) method for real applications and compare the results with other first-principle methods and experimental data, showing that they are equally as accurate.},
- keywords = {/unread,\_tablet,DFT,KKR,Poisson equation},
+ keywords = {/unread,DFT,KKR,Poisson equation},
file = {/Users/wasmer/Nextcloud/Zotero/Ning et al_2022_Full-potential KKR within the removed-sphere method.pdf}
}
@@ -12360,7 +13502,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
urldate = {2022-06-18},
isbn = {978-3-540-85416-6},
langid = {english},
- keywords = {\_tablet,condensed matter,graduate,magnetism,textbook},
+ keywords = {condensed matter,graduate,magnetism,textbook},
file = {/Users/wasmer/Nextcloud/Zotero/Quantum Theory of Magnetism.pdf;/Users/wasmer/Zotero/storage/ULV44ULF/978-3-540-85416-6.html}
}
@@ -12409,13 +13551,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Nyári, Bendegúz and Lászlóffy, András and Csire, Gábor and Szunyogh, László and Újfalussy, Balázs},
date = {2023-08-26},
eprint = {2308.13824},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2308.13824},
url = {http://arxiv.org/abs/2308.13824},
urldate = {2023-09-20},
abstract = {Magnetic chains on superconductors hosting Majorna Zero Modes (MZMs) attracted high interest due to their possible applications in fault-tolerant quantum computing. However, this is hindered by the lack of a detailed, quantitative understanding of these systems. As a significant step forward, we present a first-principles computational approach based on a microscopic relativistic theory of inhomogeneous superconductors applied to an iron chain on the top of Au-covered Nb(110) to study the Shiba band structure and the topological nature of the edge states. Contrary to contemporary considerations, our method enables the introduction of quantities indicating band inversion without fitting parameters in realistic experimental settings, holding thus the power to determine the topological nature of zero energy edge states in an accurate ab-initio based description of the experimental systems. We confirm that ferromagnetic Fe chains on Au/Nb(110) surface do not support any separated MZM; however, a broad range of spin-spirals can be identified with robust zero energy edge states displaying signatures of MZMs. For these spirals, we explore the structure of the superconducting order parameter shedding light on the internally antisymmetric triplet pairing hosted by MZMs. We also reveal a two-fold effect of spin-orbit coupling: although it tends to enlarge the topological phase regarding spin spiraling angles, however, it also extends the localization of MZMs. Due to the presented predictive power, our work fills a big gap between the experimental efforts and theoretical models while paving the way for engineering platforms for topological quantum computation.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,Budapest KKR group,GF2023 workshop,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Nyári et al_2023_Topological superconductivity from first-principles I.pdf;/Users/wasmer/Zotero/storage/A5BSXVEG/2308.html}
}
@@ -12426,13 +13568,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Ock, Janghoon and Guntuboina, Chakradhar and Farimani, Amir Barati},
date = {2023-09-01},
eprint = {2309.00563},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2309.00563},
url = {http://arxiv.org/abs/2309.00563},
urldate = {2023-11-05},
abstract = {Efficient catalyst screening necessitates predictive models for adsorption energy, a key property of reactivity. However, prevailing methods, notably graph neural networks (GNNs), demand precise atomic coordinates for constructing graph representations, while integrating observable attributes remains challenging. This research introduces CatBERTa, an energy prediction Transformer model using textual inputs. Built on a pretrained Transformer encoder, CatBERTa processes human-interpretable text, incorporating target features. Attention score analysis reveals CatBERTa's focus on tokens related to adsorbates, bulk composition, and their interacting atoms. Moreover, interacting atoms emerge as effective descriptors for adsorption configurations, while factors such as bond length and atomic properties of these atoms offer limited predictive contributions. By predicting adsorption energy from the textual representation of initial structures, CatBERTa achieves a mean absolute error (MAE) of 0.75 eV-comparable to vanilla Graph Neural Networks (GNNs). Furthermore, the subtraction of the CatBERTa-predicted energies effectively cancels out their systematic errors by as much as 19.3\% for chemically similar systems, surpassing the error reduction observed in GNNs. This outcome highlights its potential to enhance the accuracy of energy difference predictions. This research establishes a fundamental framework for text-based catalyst property prediction, without relying on graph representations, while also unveiling intricate feature-property relationships.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {adsorption,alternative approaches,alternative to GNN,AML,benchmarking,catalysis,geometric deep learning,GNN,LLM,ML,model comparison,prediction of energy,textual representation,transformer},
file = {/Users/wasmer/Nextcloud/Zotero/Ock et al_2023_Catalyst Property Prediction with CatBERTa.pdf;/Users/wasmer/Zotero/storage/E4E3G8VW/2309.html}
}
@@ -12449,7 +13591,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
doi = {10.1126/science.1237215},
url = {https://www.science.org/doi/10.1126/science.1237215},
urldate = {2022-05-13},
- keywords = {\_tablet,Hall effect,Hall QAHE,Hall QHE,Hall QSHE,perspective},
+ keywords = {for introductions,good figures,Hall effect,Hall QAHE,Hall QHE,Hall QSHE,perspective,topological,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Oh_2013_The Complete Quantum Hall Trio.pdf}
}
@@ -12504,7 +13646,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
url = {https://aip.scitation.org/doi/10.1063/5.0016005},
urldate = {2021-05-13},
abstract = {Faithfully representing chemical environments is essential for describing materials and molecules with machine learning approaches. Here, we present a systematic classification of these representations and then investigate (i) the sensitivity to perturbations and (ii) the effective dimensionality of a variety of atomic environment representations and over a range of material datasets. Representations investigated include atom centered symmetry functions, Chebyshev Polynomial Symmetry Functions (CHSF), smooth overlap of atomic positions, many-body tensor representation, and atomic cluster expansion. In area (i), we show that none of the atomic environment representations are linearly stable under tangential perturbations and that for CHSF, there are instabilities for particular choices of perturbation, which we show can be removed with a slight redefinition of the representation. In area (ii), we find that most representations can be compressed significantly without loss of precision and, further, that selecting optimal subsets of a representation method improves the accuracy of regression models built for a given dataset.},
- keywords = {\_tablet,ACE,ACSF,descriptors,MBTR,ML,MTP,SNAP,SOAP},
+ keywords = {ACE,ACSF,descriptors,MBTR,ML,MTP,SNAP,SOAP},
file = {/Users/wasmer/Nextcloud/Zotero/Onat et al_2020_Sensitivity and dimensionality of atomic environment representations used for.pdf;/Users/wasmer/Zotero/storage/RQ8UAKFX/5.html}
}
@@ -12519,20 +13661,33 @@ Subject\_term\_id: magnetic-properties-and-materials},
file = {/Users/wasmer/Zotero/storage/6TZCQAXX/Machine_Learning_For_Physicists_2021.html}
}
+@book{orlandMathProgrammers3D2020,
+ title = {Math for Programmers: {{3D}} Graphics, Machine Learning and Simulations with {{Python}}},
+ shorttitle = {Math for Programmers},
+ author = {Orland, Paul},
+ date = {2020},
+ publisher = {Manning},
+ location = {Shelter Island, NY},
+ isbn = {978-1-61729-535-5},
+ pagetotal = {655},
+ keywords = {/unread,Computer programming,Computer science,Mathematics,Python (Computer program language),Study and teaching},
+ annotation = {OCLC: on1121085014}
+}
+
@online{ortnerAtomicClusterExpansion2023,
title = {On the {{Atomic Cluster Expansion}}: Interatomic Potentials and Beyond},
shorttitle = {On the {{Atomic Cluster Expansion}}},
author = {Ortner, Christoph},
date = {2023-08-12},
eprint = {2308.06462},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.25950/c7f24234},
url = {http://arxiv.org/abs/2308.06462},
urldate = {2023-09-22},
abstract = {The Atomic Cluster Expansion (ACE) [R. Drautz, Phys. Rev. B, 99:014104 (2019)] provides a systematically improvable, universal descriptor for the environment of an atom that is invariant to permutation, translation and rotation. ACE is being used extensively in newly emerging interatomic potentials based on machine learning. This commentary discusses the ACE framework and its potential impact.},
- pubstate = {preprint},
- keywords = {\_tablet,Condensed Matter - Materials Science,Physics - Chemical Physics,Physics - Computational Physics},
+ pubstate = {prepublished},
+ keywords = {Condensed Matter - Materials Science,Physics - Chemical Physics,Physics - Computational Physics},
file = {/Users/wasmer/Zotero/storage/HD6NJWY3/Ortner_2023_On the Atomic Cluster Expansion.pdf;/Users/wasmer/Zotero/storage/LCWVPNL3/2308.html}
}
@@ -12541,13 +13696,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Ortner, Christoph and Wang, Yangshuai},
date = {2022-09-12},
eprint = {2209.05366},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math},
doi = {10.48550/arXiv.2209.05366},
url = {http://arxiv.org/abs/2209.05366},
urldate = {2023-09-22},
abstract = {Machine-learned interatomic potentials (MLIPs) and force fields (i.e. interaction laws for atoms and molecules) are typically trained on limited data-sets that cover only a very small section of the full space of possible input structures. MLIPs are nevertheless capable of making accurate predictions of forces and energies in simulations involving (seemingly) much more complex structures. In this article we propose a framework within which this kind of generalisation can be rigorously understood. As a prototypical example, we apply the framework to the case of simulating point defects in a crystalline solid. Here, we demonstrate how the accuracy of the simulation depends explicitly on the size of the training structures, on the kind of observations (e.g., energies, forces, force constants, virials) to which the model has been fitted, and on the fit accuracy. The new theoretical insights we gain partially justify current best practices in the MLIP literature and in addition suggest a new approach to the collection of training data and the design of loss functions.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,AML theory,database generation,error estimate,generalization,ML,MLP,todo-tagging},
file = {/Users/wasmer/Zotero/storage/DFJWG5I4/Ortner and Wang - 2022 - A framework for a generalisation analysis of machi.pdf;/Users/wasmer/Zotero/storage/V6S43FVE/2209.html}
}
@@ -12685,6 +13840,42 @@ Subject\_term\_id: magnetic-properties-and-materials},
file = {/Users/wasmer/Nextcloud/Zotero/Pant et al_2023_DFT-aided machine learning-based discovery of magnetism in Fe-based bimetallic.pdf}
}
+@online{papamarkouPositionPaperBayesian2024,
+ title = {Position {{Paper}}: {{Bayesian Deep Learning}} in the {{Age}} of {{Large-Scale AI}}},
+ shorttitle = {Position {{Paper}}},
+ author = {Papamarkou, Theodore and Skoularidou, Maria and Palla, Konstantina and Aitchison, Laurence and Arbel, Julyan and Dunson, David and Filippone, Maurizio and Fortuin, Vincent and Hennig, Philipp and Lobato, Jose Miguel Hernandez and Hubin, Aliaksandr and Immer, Alexander and Karaletsos, Theofanis and Khan, Mohammad Emtiyaz and Kristiadi, Agustinus and Li, Yingzhen and Mandt, Stephan and Nemeth, Christopher and Osborne, Michael A. and Rudner, Tim G. J. and Rügamer, David and Teh, Yee Whye and Welling, Max and Wilson, Andrew Gordon and Zhang, Ruqi},
+ date = {2024-02-06},
+ eprint = {2402.00809},
+ eprinttype = {arXiv},
+ eprintclass = {cs, stat},
+ doi = {10.48550/arXiv.2402.00809},
+ url = {http://arxiv.org/abs/2402.00809},
+ urldate = {2024-05-27},
+ abstract = {In the current landscape of deep learning research, there is a predominant emphasis on achieving high predictive accuracy in supervised tasks involving large image and language datasets. However, a broader perspective reveals a multitude of overlooked metrics, tasks, and data types, such as uncertainty, active and continual learning, and scientific data, that demand attention. Bayesian deep learning (BDL) constitutes a promising avenue, offering advantages across these diverse settings. This paper posits that BDL can elevate the capabilities of deep learning. It revisits the strengths of BDL, acknowledges existing challenges, and highlights some exciting research avenues aimed at addressing these obstacles. Looking ahead, the discussion focuses on possible ways to combine large-scale foundation models with BDL to unlock their full potential.},
+ pubstate = {prepublished},
+ keywords = {/unread,AI4Science,Bayesian deep learning,Bayesian methods,Deep learning,ensemble learning,foundation models,General ML,hallucinations,large models,LLM,ML,overconfidence,position paper,roadmap,XAI},
+ file = {/Users/wasmer/Nextcloud/Zotero/Papamarkou et al_2024_Position Paper.pdf;/Users/wasmer/Zotero/storage/JLX9YRAV/2402.html}
+}
+
+@article{parkScalableParallelAlgorithm2024,
+ title = {Scalable {{Parallel Algorithm}} for {{Graph Neural Network Interatomic Potentials}} in {{Molecular Dynamics Simulations}}},
+ author = {Park, Yutack and Kim, Jaesun and Hwang, Seungwoo and Han, Seungwu},
+ date = {2024-06-11},
+ journaltitle = {Journal of Chemical Theory and Computation},
+ shortjournal = {J. Chem. Theory Comput.},
+ volume = {20},
+ number = {11},
+ pages = {4857--4868},
+ publisher = {American Chemical Society},
+ issn = {1549-9618},
+ doi = {10.1021/acs.jctc.4c00190},
+ url = {https://doi.org/10.1021/acs.jctc.4c00190},
+ urldate = {2024-07-17},
+ abstract = {Message-passing graph neural network interatomic potentials (GNN-IPs), particularly those with equivariant representations such as NequIP, are attracting significant attention due to their data efficiency and high accuracy. However, parallelizing GNN-IPs poses challenges because multiple message-passing layers complicate data communication within the spatial decomposition method, which is preferred by many molecular dynamics (MD) packages. In this article, we propose an efficient parallelization scheme compatible with GNN-IPs and develop a package, SevenNet (Scalable EquiVariance-Enabled Neural NETwork), based on the NequIP architecture. For MD simulations, SevenNet interfaces with the LAMMPS package. Through benchmark tests on a 32-GPU cluster with examples of SiO2, SevenNet achieves over 80\% parallel efficiency in weak-scaling scenarios and exhibits nearly ideal strong-scaling performance as long as GPUs are fully utilized. However, the strong-scaling performance significantly declines with suboptimal GPU utilization, particularly affecting parallel efficiency in cases involving lightweight models or simulations with small numbers of atoms. We also pretrain SevenNet with a vast data set from the Materials Project (dubbed “SevenNet-0â€) and assess its performance on generating amorphous Si3N4 containing more than 100,000 atoms. By developing scalable GNN-IPs, this work aims to bridge the gap between advanced machine-learning models and large-scale MD simulations, offering researchers a powerful tool to explore complex material systems with high accuracy and efficiency.},
+ keywords = {/unread,AML,HPC,hybrid AI/simulation,integrated models,LAMMPS,large models,large simulations,library,MatBench,materials project,MD,ML,ML-IAP,NequIP,Parallel computing,parallelization,strong scaling,supercomputing,universal potential,weak scaling,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Park et al_2024_Scalable Parallel Algorithm for Graph Neural Network Interatomic Potentials in.pdf}
+}
+
@article{parsaeifardAssessmentStructuralResolution2021,
title = {An Assessment of the Structural Resolution of Various Fingerprints Commonly Used in Machine Learning},
author = {Parsaeifard, Behnam and De, Deb Sankar and Christensen, Anders S. and Faber, Felix A. and Kocer, Emir and De, Sandip and Behler, Jörg and family=Lilienfeld, given=O. Anatole, prefix=von, useprefix=false and Goedecker, Stefan},
@@ -12793,7 +13984,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
urldate = {2023-09-19},
abstract = {We developed a graph convolutional neural network (GCNN) to predict the formation energy and the bulk modulus for models of solid solution alloys for various atomic crystal structures and relaxed volumes. We trained the GCNN model on a dataset for nickel-niobium (NiNb) that was generated for simplicity with the embedded atom model (EAM) empirical interatomic potential. The dataset has been generated by calculating the formation energy and the bulk modulus as a prototypical elastic property for optimized geometries starting from initial body-centered cubic (BCC), face-centered cubic (FCC), and hexagonal compact packed (HCP) crystal structures, with configurations spanning the possible compositional range for each of the three types of initial crystal structure. Numerical results show that the GCNN model effectively predicts both the formation energy and the bulk modulus as functions of the optimized crystal structure, relaxed volume, and configurational entropy of the model structures for solid solution alloys.},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Pasini et al_2022_Graph neural networks predict energetic and mechanical properties for models of.pdf}
}
@@ -12814,7 +14005,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
urldate = {2023-09-19},
abstract = {We introduce a multi-tasking graph convolutional neural network, HydraGNN, to simultaneously predict both global and atomic physical properties and demonstrate with ferromagnetic materials. We train HydraGNN on an open-source ab initio density functional theory (DFT) dataset for iron-platinum with a fixed body centered tetragonal lattice structure and fixed volume to simultaneously predict the mixing enthalpy (a global feature of the system), the atomic charge transfer, and the atomic magnetic moment across configurations that span the entire compositional range. By taking advantage of underlying physical correlations between material properties, multi-task learning (MTL) with HydraGNN provides effective training even with modest amounts of data. Moreover, this is achieved with just one architecture instead of three, as required by single-task learning (STL). The first convolutional layers of the HydraGNN architecture are shared by all learning tasks and extract features common to all material properties. The following layers discriminate the features of the different properties, the results of which are fed to the separate heads of the final layer to produce predictions. Numerical results show that HydraGNN effectively captures the relation between the configurational entropy and the material properties over the entire compositional range. Overall, the accuracy of simultaneous MTL predictions is comparable to the accuracy of the STL predictions. In addition, the computational cost of training HydraGNN for MTL is much lower than the original DFT calculations and also lower than training separate STL models for each property.},
langid = {english},
- keywords = {\_tablet,AML,binary systems,charge transfer,DFT,Ferromagnetism,GF2023 workshop,GNN,HydraGNN,library,magnetism,ML,ML-DFT,ML-ESM,multi-task learning,ORNL,prediction of charge transfer,prediction of energy,prediction of magnetic moment,PyTorch,surrogate model,with-code,with-data},
+ keywords = {AML,binary systems,charge transfer,DFT,Ferromagnetism,GF2023 workshop,GNN,HydraGNN,library,magnetism,ML,ML-DFT,ML-ESM,multi-task learning,ORNL,prediction of charge transfer,prediction of energy,prediction of magnetic moment,PyTorch,surrogate model,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Pasini et al_2022_Multi-task graph neural networks for simultaneous prediction of global and.pdf}
}
@@ -12828,7 +14019,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
urldate = {2023-09-19},
abstract = {In this study, we show the transferability of graph convolutional neural network (GCNN) predictions of the formation energy of solid solution alloys across atomic structures of increasing sizes, which was utilized in the cost-efficient sampling strategy. The GCNN was trained on a nickel-platinum (NiPt) dataset generated with the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) using the second nearest-neighbor modified embedded-atom method (2NN MEAM) empirical interatomic potential. The dataset has been obtained by optimizing the geometries of initially randomly generated FCC crystal structures and calculating the formation energy, with configurations spanning the whole compositional range. The GCNN was first trained on a lattice of 256 atoms, which accounts well for the short-range interactions. Using this data, we predicted the formation energy for lattices of 864 atoms and 2,048 atoms, which resulted in lower-than-expected accuracy due to the long-range interactions present in these larger lattices. We accounted for the long-range interactions by including a small amount of training data representative for those two larger sizes. Using this additional data, the predictions of the GCNN scaled linearly with the size of the lattice. Therefore, our strategy ensured scalability while reducing significantly the computational cost of training on larger lattice sizes.},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {HydraGNN,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Pasini et al_2023_Transferable prediction of formation energy across lattices of increasing size.pdf}
}
@@ -12838,13 +14029,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Passaro, Saro and Zitnick, C. Lawrence},
date = {2023-06-14},
eprint = {2302.03655},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2302.03655},
url = {http://arxiv.org/abs/2302.03655},
urldate = {2024-05-07},
abstract = {Graph neural networks that model 3D data, such as point clouds or atoms, are typically desired to be \$SO(3)\$ equivariant, i.e., equivariant to 3D rotations. Unfortunately equivariant convolutions, which are a fundamental operation for equivariant networks, increase significantly in computational complexity as higher-order tensors are used. In this paper, we address this issue by reducing the \$SO(3)\$ convolutions or tensor products to mathematically equivalent convolutions in \$SO(2)\$ . This is accomplished by aligning the node embeddings' primary axis with the edge vectors, which sparsifies the tensor product and reduces the computational complexity from \$O(L\textasciicircum 6)\$ to \$O(L\textasciicircum 3)\$, where \$L\$ is the degree of the representation. We demonstrate the potential implications of this improvement by proposing the Equivariant Spherical Channel Network (eSCN), a graph neural network utilizing our novel approach to equivariant convolutions, which achieves state-of-the-art results on the large-scale OC-20 and OC-22 datasets.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,alternative approaches,alternative for equivariance,AML,computational complexity,convolution,equivariant,equivariant alternative,eSCN,GNN,Meta Research,ML,MLP,MPNN,Open Catalyst,rotational symmetry,SO(3),tensor product,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Passaro_Zitnick_2023_Reducing SO(3) Convolutions to SO(2) for Efficient Equivariant GNNs2.pdf;/Users/wasmer/Zotero/storage/IIL5PCZ5/2302.html}
}
@@ -12916,13 +14107,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Peach, Robert L. and Vinao-Carl, Matteo and Grossman, Nir and David, Michael and Mallas, Emma and Sharp, David and Malhotra, Paresh A. and Vandergheynst, Pierre and Gosztolai, Adam},
date = {2023-09-28},
eprint = {2309.16746},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, q-bio, stat},
doi = {10.48550/arXiv.2309.16746},
url = {http://arxiv.org/abs/2309.16746},
urldate = {2023-10-07},
abstract = {Gaussian processes (GPs) are popular nonparametric statistical models for learning unknown functions and quantifying the spatiotemporal uncertainty in data. Recent works have extended GPs to model scalar and vector quantities distributed over non-Euclidean domains, including smooth manifolds appearing in numerous fields such as computer vision, dynamical systems, and neuroscience. However, these approaches assume that the manifold underlying the data is known, limiting their practical utility. We introduce RVGP, a generalisation of GPs for learning vector signals over latent Riemannian manifolds. Our method uses positional encoding with eigenfunctions of the connection Laplacian, associated with the tangent bundle, readily derived from common graph-based approximation of data. We demonstrate that RVGP possesses global regularity over the manifold, which allows it to super-resolve and inpaint vector fields while preserving singularities. Furthermore, we use RVGP to reconstruct high-density neural dynamics derived from low-density EEG recordings in healthy individuals and Alzheimer's patients. We show that vector field singularities are important disease markers and that their reconstruction leads to a comparable classification accuracy of disease states to high-density recordings. Thus, our method overcomes a significant practical limitation in experimental and clinical applications.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,Gaussian process,General ML,library,Manifolds,ML,singularities,tensorial target,vector field,vectorial learning target,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Peach et al_2023_Implicit Gaussian process representation of vector fields over arbitrary latent.pdf;/Users/wasmer/Zotero/storage/J7N49F7L/2309.html}
}
@@ -12932,7 +14123,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Pederson, Ryan and Kalita, Bhupalee and Burke, Kieron},
date = {2022-05-03},
eprint = {2205.01591},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2205.01591},
urldate = {2022-05-13},
@@ -12947,7 +14138,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Pedregosa, Fabian and Varoquaux, Gaël and Gramfort, Alexandre and Michel, Vincent and Thirion, Bertrand and Grisel, Olivier and Blondel, Mathieu and Müller, Andreas and Nothman, Joel and Louppe, Gilles and Prettenhofer, Peter and Weiss, Ron and Dubourg, Vincent and Vanderplas, Jake and Passos, Alexandre and Cournapeau, David and Brucher, Matthieu and Perrot, Matthieu and Duchesnay, Édouard},
date = {2018-06-05},
eprint = {1201.0490},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
url = {http://arxiv.org/abs/1201.0490},
urldate = {2021-07-14},
@@ -12977,23 +14168,6 @@ Subject\_term\_id: magnetic-properties-and-materials},
file = {/Users/wasmer/Nextcloud/Zotero/Peixoto et al_2020_Non-local effect of impurity states on the exchange coupling mechanism in.pdf;/Users/wasmer/Zotero/storage/DDIQNTSB/s41535-020-00288-0.html}
}
-@online{penzStructureDensitypotentialMapping2023,
- title = {The Structure of the Density-Potential Mapping. {{Part II}}: {{Including}} Magnetic Fields},
- shorttitle = {The Structure of the Density-Potential Mapping. {{Part II}}},
- author = {Penz, Markus and Tellgren, Erik I. and Csirik, Mihály A. and Ruggenthaler, Michael and Laestadius, Andre},
- date = {2023-03-02},
- eprint = {2303.01357},
- eprinttype = {arxiv},
- eprintclass = {physics, physics:quant-ph},
- doi = {10.48550/arXiv.2303.01357},
- url = {http://arxiv.org/abs/2303.01357},
- urldate = {2023-06-30},
- abstract = {The Hohenberg-Kohn theorem of density-functional theory (DFT) is broadly considered the conceptual basis for a full characterization of an electronic system in its ground state by just the one-body particle density. In this Part\textasciitilde II of a series of two articles, we aim at clarifying the status of this theorem within different extensions of DFT including magnetic fields. We will in particular discuss current-density-functional theory (CDFT) and review the different formulations known in the literature, including the conventional paramagnetic CDFT and some non-standard alternatives. For the former, it is known that the Hohenberg-Kohn theorem is no longer valid due to counterexamples. Nonetheless, paramagnetic CDFT has the mathematical framework closest to standard DFT and, just like in standard DFT, non-differentiability of the density functional can be mitigated through Moreau-Yosida regularization. Interesting insights can be drawn from both Maxwell-Schr\textbackslash "odinger DFT and quantum-electrodynamical DFT, which are also discussed here.},
- pubstate = {preprint},
- keywords = {DFT,DFT theory,HK map,HKT,magnetism,Physics - Chemical Physics,Quantum Physics,review,review-of-DFT},
- file = {/Users/wasmer/Nextcloud/Zotero/Penz et al_2023_The structure of the density-potential mapping.pdf;/Users/wasmer/Zotero/storage/G52G7MTD/2303.html}
-}
-
@article{penzStructureDensityPotentialMapping2023,
title = {The {{Structure}} of {{Density-Potential Mapping}}. {{Part I}}: {{Standard Density-Functional Theory}}},
shorttitle = {The {{Structure}} of {{Density-Potential Mapping}}. {{Part I}}},
@@ -13010,6 +14184,23 @@ Subject\_term\_id: magnetic-properties-and-materials},
file = {/Users/wasmer/Nextcloud/Zotero/Penz et al_2023_The Structure of Density-Potential Mapping.pdf;/Users/wasmer/Zotero/storage/ASJHHVMZ/acsphyschemau.html}
}
+@online{penzStructureDensitypotentialMapping2023a,
+ title = {The Structure of the Density-Potential Mapping. {{Part II}}: {{Including}} Magnetic Fields},
+ shorttitle = {The Structure of the Density-Potential Mapping. {{Part II}}},
+ author = {Penz, Markus and Tellgren, Erik I. and Csirik, Mihály A. and Ruggenthaler, Michael and Laestadius, Andre},
+ date = {2023-03-02},
+ eprint = {2303.01357},
+ eprinttype = {arXiv},
+ eprintclass = {physics, physics:quant-ph},
+ doi = {10.48550/arXiv.2303.01357},
+ url = {http://arxiv.org/abs/2303.01357},
+ urldate = {2023-06-30},
+ abstract = {The Hohenberg-Kohn theorem of density-functional theory (DFT) is broadly considered the conceptual basis for a full characterization of an electronic system in its ground state by just the one-body particle density. In this Part\textasciitilde II of a series of two articles, we aim at clarifying the status of this theorem within different extensions of DFT including magnetic fields. We will in particular discuss current-density-functional theory (CDFT) and review the different formulations known in the literature, including the conventional paramagnetic CDFT and some non-standard alternatives. For the former, it is known that the Hohenberg-Kohn theorem is no longer valid due to counterexamples. Nonetheless, paramagnetic CDFT has the mathematical framework closest to standard DFT and, just like in standard DFT, non-differentiability of the density functional can be mitigated through Moreau-Yosida regularization. Interesting insights can be drawn from both Maxwell-Schr\textbackslash "odinger DFT and quantum-electrodynamical DFT, which are also discussed here.},
+ pubstate = {prepublished},
+ keywords = {DFT,DFT theory,HK map,HKT,magnetism,Physics - Chemical Physics,Quantum Physics,review,review-of-DFT},
+ file = {/Users/wasmer/Nextcloud/Zotero/Penz et al_2023_The structure of the density-potential mapping.pdf;/Users/wasmer/Zotero/storage/G52G7MTD/2303.html}
+}
+
@article{pereiraChallengesTopologicalInsulator2021,
title = {Challenges of {{Topological Insulator Research}}: {{Bi2Te3 Thin Films}} and {{Magnetic Heterostructures}}},
shorttitle = {Challenges of {{Topological Insulator Research}}},
@@ -13035,13 +14226,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Persaud, Daniel and Ward, Logan and Hattrick-Simpers, Jason},
date = {2023-10-10},
eprint = {2310.07044},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2310.07044},
url = {http://arxiv.org/abs/2310.07044},
urldate = {2023-12-05},
abstract = {The integration of machine learning techniques in materials discovery has become prominent in materials science research and has been accompanied by an increasing trend towards open-source data and tools to propel the field. Despite the increasing usefulness and capabilities of these tools, developers neglecting to follow reproducible practices creates a significant barrier for researchers looking to use or build upon their work. In this study, we investigate the challenges encountered while attempting to reproduce a section of the results presented in "A general-purpose machine learning framework for predicting properties of inorganic materials." Our analysis identifies four major categories of challenges: (1) reporting computational dependencies, (2) recording and sharing version logs, (3) sequential code organization, and (4) clarifying code references within the manuscript. The result is a proposed set of tangible action items for those aiming to make code accessible to, and useful for the community.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,materials informatics,metadata,ML,RDM,reproducibility,scientific workflows,todo-tagging,version control,workflows},
file = {/Users/wasmer/Nextcloud/Zotero/Persaud et al_2023_Reproducibility in Computational Materials Science.pdf;/Users/wasmer/Zotero/storage/XPWR5SBW/2310.html}
}
@@ -13073,7 +14264,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
volume = {2},
number = {3},
doi = {10.1103/PhysRevResearch.2.033429},
- keywords = {DeepMind,FermiNet,JAX,library,MC,ML,ML-ESM,ML-QMBP,NN,PauliNet,prediction of wavefunction,QMC,VMC,with-code},
+ keywords = {DeepMind,FermiNet,JAX,library,MC,ML,ML-ESM,ML-QMBP,NN,prediction of wavefunction,QMC,VMC,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Pfau_2020_iAb initio-i solution of the many-electron Schrödinger equation with deep.pdf;/Users/wasmer/Zotero/storage/7HFHVNYZ/PhysRevResearch.2.html}
}
@@ -13113,13 +14304,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Pinheiro, Pedro O. and Rackers, Joshua and Kleinhenz, Joseph and Maser, Michael and Mahmood, Omar and Watkins, Andrew Martin and Ra, Stephen and Sresht, Vishnu and Saremi, Saeed},
date = {2023-06-12},
eprint = {2306.07473},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, q-bio},
doi = {10.48550/arXiv.2306.07473},
url = {http://arxiv.org/abs/2306.07473},
urldate = {2023-08-19},
abstract = {We propose a new score-based approach to generate 3D molecules represented as atomic densities on regular grids. First, we train a denoising neural network that learns to map from a smooth distribution of noisy molecules to the distribution of real molecules. Then, we follow the neural empirical Bayes framework [Saremi and Hyvarinen, 2019] and generate molecules in two steps: (i) sample noisy density grids from a smooth distribution via underdamped Langevin Markov chain Monte Carlo, and (ii) recover the ``clean'' molecule by denoising the noisy grid with a single step. Our method, VoxMol, generates molecules in a fundamentally different way than the current state of the art (i.e., diffusion models applied to atom point clouds). It differs in terms of the data representation, the noise model, the network architecture and the generative modeling algorithm. VoxMol achieves comparable results to state of the art on unconditional 3D molecule generation while being simpler to train and faster to generate molecules.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Pinheiro et al_2023_3D molecule generation by denoising voxel grids.pdf;/Users/wasmer/Zotero/storage/6CKBVABI/2306.html}
}
@@ -13186,13 +14377,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Podryabinkin, Evgeny and Garifullin, Kamil and Shapeev, Alexander and Novikov, Ivan},
date = {2023-04-25},
eprint = {2304.13144},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2304.13144},
url = {http://arxiv.org/abs/2304.13144},
urldate = {2023-05-26},
abstract = {Nowadays, academic research relies not only on sharing with the academic community the scientific results obtained by research groups while studying certain phenomena, but also on sharing computer codes developed within the community. In the field of atomistic modeling these were software packages for classical atomistic modeling, later -- quantum-mechanical modeling, and now with the fast growth of the field of machine-learning potentials, the packages implementing such potentials. In this paper we present the MLIP-3 package for constructing moment tensor potentials and performing their active training. This package builds on the MLIP-2 package (Novikov et al. (2020), The MLIP package: moment tensor potentials with MPI and active learning. Machine Learning: Science and Technology, 2(2), 025002.), however with a number of improvements, including active learning on atomic neighborhoods of a possibly large atomistic simulation.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {active learning,active learning online,AML,library,ML,MLP,MTP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Podryabinkin et al_2023_MLIP-3.pdf;/Users/wasmer/Zotero/storage/QTPXSGSW/2304.html}
}
@@ -13233,7 +14424,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
urldate = {2023-03-19},
abstract = {We introduce a machine-learning (ML) framework for high-throughput benchmarking of diverse representations of chemical systems against datasets of materials and molecules. The guiding principle underlying the benchmarking approach is to evaluate raw descriptor performance by limiting model complexity to simple regression schemes while enforcing best ML practices, allowing for unbiased hyperparameter optimization, and assessing learning progress through learning curves along series of synchronized train-test splits. The resulting models are intended as baselines that can inform future method development, in addition to indicating how easily a given dataset can be learnt. Through a comparative analysis of the training outcome across a diverse set of physicochemical, topological and geometric representations, we glean insight into the relative merits of these representations as well as their interrelatedness.},
langid = {english},
- keywords = {\_tablet,ACSF,AML,benchmarking,CM,Coulomb matrix,descriptor comparison,descriptors,DScribe,ECFP descriptor,GYLM descriptor,KRR,library,linear regression,materials,MBTR,ML,MLOps,molecules,QM7b,QM9,SISSO,SOAP,with-code,workflows},
+ keywords = {ACSF,AML,benchmarking,CM,Coulomb matrix,descriptor comparison,descriptors,DScribe,ECFP descriptor,GYLM descriptor,KRR,library,linear regression,materials,MBTR,ML,MLOps,molecules,QM7b,QM9,SISSO,SOAP,with-code,workflows},
file = {/Users/wasmer/Zotero/storage/QXAEL2PM/Poelking et al_2022_BenchML.pdf}
}
@@ -13260,13 +14451,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Polak, Maciej P. and Morgan, Dane},
date = {2023-03-07},
eprint = {2303.05352},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2303.05352},
url = {http://arxiv.org/abs/2303.05352},
urldate = {2023-03-17},
abstract = {There has been a growing effort to replace hand extraction of data from research papers with automated data extraction based on natural language processing (NLP), language models (LMs), and recently, large language models (LLMs). Although these methods enable efficient extraction of data from large sets of research papers, they require a significant amount of up-front effort, expertise, and coding. In this work we propose the ChatExtract method that can fully automate very accurate data extraction with essentially no initial effort or background using an advanced conversational LLM (or AI). ChatExtract consists of a set of engineered prompts applied to a conversational LLM that both identify sentences with data, extract data, and assure its correctness through a series of follow-up questions. These follow-up questions address a critical challenge associated with LLMs - their tendency to provide factually inaccurate responses. ChatExtract can be applied with any conversational LLMs and yields very high quality data extraction. In tests on materials data we find precision and recall both over 90\% from the best conversational LLMs, likely rivaling or exceeding human accuracy in many cases. We demonstrate that the exceptional performance is enabled by the information retention in a conversational model combined with purposeful redundancy and introducing uncertainty through follow-up prompts. These results suggest that approaches similar to ChatExtract, due to their simplicity, transferability and accuracy are likely to replace other methods of data extraction in the near future.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,ChatGPT,data mining,database generation,GPT,GPT-3,literature analysis,LLM,materials,prompt engineering},
file = {/Users/wasmer/Nextcloud/Zotero/Polak_Morgan_2023_Extracting Accurate Materials Data from Research Papers with Conversational.pdf;/Users/wasmer/Zotero/storage/9HJ5N4FB/2303.html}
}
@@ -13276,17 +14467,35 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Polak, Maciej P. and Modi, Shrey and Latosinska, Anna and Zhang, Jinming and Wang, Ching-Wen and Wang, Shanonan and Hazra, Ayan Deep and Morgan, Dane},
date = {2023-02-09},
eprint = {2302.04914},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2302.04914},
url = {http://arxiv.org/abs/2302.04914},
urldate = {2023-03-17},
abstract = {Accurate and comprehensive material databases extracted from research papers are critical for materials science and engineering but require significant human effort to develop. In this paper we present a simple method of extracting materials data from full texts of research papers suitable for quickly developing modest-sized databases. The method requires minimal to no coding, prior knowledge about the extracted property, or model training, and provides high recall and almost perfect precision in the resultant database. The method is fully automated except for one human-assisted step, which typically requires just a few hours of human labor. The method builds on top of natural language processing and large general language models but can work with almost any such model. The language models GPT-3/3.5, bart and DeBERTaV3 are evaluated here for comparison. We provide a detailed detailed analysis of the methods performance in extracting bulk modulus data, obtaining up to 90\% precision at 96\% recall, depending on the amount of human effort involved. We then demonstrate the methods broader effectiveness by developing a database of critical cooling rates for metallic glasses.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ChatGPT,data mining,database generation,GPT,GPT-3,literature analysis,LLM,materials},
file = {/Users/wasmer/Nextcloud/Zotero/Polak et al_2023_Flexible, Model-Agnostic Method for Materials Data Extraction from Text Using.pdf;/Users/wasmer/Zotero/storage/2BSBWMWC/2302.html}
}
+@article{polashTopologicalQuantumMatter2021,
+ title = {Topological Quantum Matter to Topological Phase Conversion: {{Fundamentals}}, Materials, Physical Systems for Phase Conversions, and Device Applications},
+ shorttitle = {Topological Quantum Matter to Topological Phase Conversion},
+ author = {Polash, Md Mobarak Hossain and Yalameha, Shahram and Zhou, Haihan and Ahadi, Kaveh and Nourbakhsh, Zahra and Vashaee, Daryoosh},
+ date = {2021-07-01},
+ journaltitle = {Materials Science and Engineering: R: Reports},
+ shortjournal = {Materials Science and Engineering: R: Reports},
+ volume = {145},
+ pages = {100620},
+ issn = {0927-796X},
+ doi = {10.1016/j.mser.2021.100620},
+ url = {https://www.sciencedirect.com/science/article/pii/S0927796X21000152},
+ urldate = {2024-05-19},
+ abstract = {The spin-orbit coupling field, an atomic magnetic field inside a Kramers’ system, or discrete symmetries can create a topological torus in the Brillouin Zone and provide protected edge or surface states, which can contain relativistic fermions, namely, Dirac and Weyl Fermions. The topology-protected helical edge or surface states and the bulk electronic energy band define different quantum or topological phases of matters, offering an excellent prospect for some unique device applications. Device applications of the quantum materials rely primarily on understanding the topological properties, their mutual conversion processes under different external stimuli, and the physical system for achieving the phase conversion. There have been tremendous efforts in finding new topological materials with exotic topological phases. However, the application of the topological properties in devices is still limited due to the slow progress in developing the physical structures for controlling the topological phase conversions. Such control systems often require extreme tuning conditions or the fabrication of complex multi-layered topological structures. This review article highlights the details of the topological phases, their conversion processes, along with their potential physical systems, and the prospective application fields. A general overview of the critical factors for topological phases and the materials properties are further discussed to provide the necessary background for the following sections.},
+ keywords = {condensed matter,good figures,physics,quantum materials,review,review-of-TIs,topological,topological insulator,topological phase,topological phase transition},
+ file = {/Users/wasmer/Nextcloud/Zotero/Polash et al_2021_Topological quantum matter to topological phase conversion2.pdf;/Users/wasmer/Zotero/storage/KY4YEXRB/S0927796X21000152.html}
+}
+
@online{PossibleGameChanger,
title = {A Possible Game Changer for next Generation Microelectronics | {{Argonne National Laboratory}}},
url = {https://www.anl.gov/article/a-possible-game-changer-for-next-generation-microelectronics},
@@ -13328,7 +14537,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
url = {https://link.aps.org/doi/10.1103/PhysRevLett.125.166001},
urldate = {2021-05-13},
abstract = {Many-body descriptors are widely used to represent atomic environments in the construction of machine-learned interatomic potentials and more broadly for fitting, classification, and embedding tasks on atomic structures. There is a widespread belief in the community that three-body correlations are likely to provide an overcomplete description of the environment of an atom. We produce several counterexamples to this belief, with the consequence that any classifier, regression, or embedding model for atom-centered properties that uses three- (or four)-body features will incorrectly give identical results for different configurations. Writing global properties (such as total energies) as a sum of many atom-centered contributions mitigates the impact of this fundamental deficiency—explaining the success of current “machine-learning†force fields. We anticipate the issues that will arise as the desired accuracy increases, and suggest potential solutions.},
- keywords = {\_tablet,3-body order descriptors,descriptors,descriptors analysis,GPR,incompleteness,MBTR,ML,SOAP},
+ keywords = {3-body order descriptors,descriptors,descriptors analysis,GPR,incompleteness,MBTR,ML,SOAP},
file = {/Users/wasmer/Nextcloud/Zotero/Pozdnyakov et al_2020_Incompleteness of Atomic Structure Representations.pdf;/Users/wasmer/Zotero/storage/5QHMC4CR/PhysRevLett.125.html}
}
@@ -13350,17 +14559,33 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Pozdnyakov, Sergey N. and Ceriotti, Michele},
date = {2023-05-30},
eprint = {2305.19302},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2305.19302},
url = {http://arxiv.org/abs/2305.19302},
urldate = {2023-06-01},
abstract = {Point clouds are versatile representations of 3D objects and have found widespread application in science and engineering. Many successful deep-learning models have been proposed that use them as input. Some application domains require incorporating exactly physical constraints, including chemical and materials modeling which we focus on in this paper. These constraints include smoothness, and symmetry with respect to translations, rotations, and permutations of identical particles. Most existing architectures in other domains do not fulfill simultaneously all of these requirements and thus are not applicable to atomic-scale simulations. Many of them, however, can be straightforwardly made to incorporate all the physical constraints except for rotational symmetry. We propose a general symmetrization protocol that adds rotational equivariance to any given model while preserving all the other constraints. As a demonstration of the potential of this idea, we introduce the Point Edge Transformer (PET) architecture, which is not intrinsically equivariant but achieves state-of-the-art performance on several benchmark datasets of molecules and solids. A-posteriori application of our general protocol makes PET exactly equivariant, with minimal changes to its accuracy. By alleviating the need to explicitly incorporate rotational symmetry within the model, our method bridges the gap between the approaches used in different communities, and simplifies the design of deep-learning schemes for chemical and materials modeling.},
- pubstate = {preprint},
- keywords = {\_tablet,alternative approaches,alternative to GNN,AML,approximative equivariance,chemical species scaling problem,collinear,descriptors,equivariant,equivariant alternative,GNN,MACE,ML,MPNN,NequIP,point cloud data,representation learning,rotational symmetry,simplification,SO(3),spin-dependent,spin-polarized,symmetrization,symmetry,transformer,with-code},
+ pubstate = {prepublished},
+ keywords = {alternative approaches,alternative to GNN,AML,approximative equivariance,chemical species scaling problem,collinear,descriptors,equivariant,equivariant alternative,GNN,MACE,ML,MPNN,NequIP,point cloud data,representation learning,rotational symmetry,simplification,SO(3),spin-dependent,spin-polarized,symmetrization,symmetry,transformer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Pozdnyakov_Ceriotti_2023_Smooth, exact rotational symmetrization for deep learning on point clouds.pdf;/Users/wasmer/Zotero/storage/W32HXDSQ/2305.html}
}
+@online{pozdnyakovSmoothExactRotational2024,
+ title = {Smooth, Exact Rotational Symmetrization for Deep Learning on Point Clouds},
+ author = {Pozdnyakov, Sergey N. and Ceriotti, Michele},
+ date = {2024-02-06},
+ eprint = {2305.19302},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat, physics:physics},
+ doi = {10.48550/arXiv.2305.19302},
+ url = {http://arxiv.org/abs/2305.19302},
+ urldate = {2024-06-27},
+ abstract = {Point clouds are versatile representations of 3D objects and have found widespread application in science and engineering. Many successful deep-learning models have been proposed that use them as input. The domain of chemical and materials modeling is especially challenging because exact compliance with physical constraints is highly desirable for a model to be usable in practice. These constraints include smoothness and invariance with respect to translations, rotations, and permutations of identical atoms. If these requirements are not rigorously fulfilled, atomistic simulations might lead to absurd outcomes even if the model has excellent accuracy. Consequently, dedicated architectures, which achieve invariance by restricting their design space, have been developed. General-purpose point-cloud models are more varied but often disregard rotational symmetry. We propose a general symmetrization method that adds rotational equivariance to any given model while preserving all the other requirements. Our approach simplifies the development of better atomic-scale machine-learning schemes by relaxing the constraints on the design space and making it possible to incorporate ideas that proved effective in other domains. We demonstrate this idea by introducing the Point Edge Transformer (PET) architecture, which is not intrinsically equivariant but achieves state-of-the-art performance on several benchmark datasets of molecules and solids. A-posteriori application of our general protocol makes PET exactly equivariant, with minimal changes to its accuracy.},
+ pubstate = {prepublished},
+ keywords = {/unread,alternative approaches,alternative to GNN,AML,approximative equivariance,chemical species scaling problem,collinear,descriptors,equivariant,equivariant alternative,GNN,MACE,ML,MPNN,NequIP,NeurIPS,point cloud data,representation learning,rotational symmetry,simplification,SO(3),spin-dependent,spin-polarized,symmetrization,symmetry,transformer,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Pozdnyakov_Ceriotti_2024_Smooth, exact rotational symmetrization for deep learning on point clouds.pdf;/Users/wasmer/Zotero/storage/X8KQDF53/2305.html}
+}
+
@book{princeUnderstandingDeepLearning2023,
title = {Understanding Deep Learning},
author = {Prince, Simon J. D.},
@@ -13371,7 +14596,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
abstract = {"This book covers modern deep learning and tackles supervised learning, model architecture, unsupervised learning, and deep reinforcement learning"--},
isbn = {978-0-262-37709-6 978-0-262-37710-2},
pagetotal = {1},
- keywords = {\_tablet,Deep learning,educational,GNN,graph ML,learning material,ML theory,online book,online course,textbook,transformer,tutorial,with-code},
+ keywords = {Deep learning,educational,GNN,graph ML,learning material,ML theory,online book,online course,textbook,transformer,tutorial,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Prince_2023_Understanding deep learning.pdf}
}
@@ -13386,7 +14611,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
urldate = {2023-04-10},
abstract = {While the introduction of practical deep learning has driven progress across scientific fields, recent research highlighted that deep learning has potential negative impacts on the scientific community and society as a whole. An ever-growing need for more computational resources may exacerbate the concentration of funding and the exclusiveness of research between countries, sectors, and institutions. Here, I introduce recent concerns and considerations of the machine learning research community and present potential solutions.},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,best practices,chemistry,computational cost,cost analysis,criticism,Deep learning,LLM,ML,ML cost analysis,ML ethics,model evaluation,model reporting,skepticism,small data},
file = {/Users/wasmer/Nextcloud/Zotero/Probst_2022_Growing pains.pdf}
}
@@ -13407,18 +14632,38 @@ Subject\_term\_id: magnetic-properties-and-materials},
file = {/Users/wasmer/Nextcloud/Zotero/Prodan_Kohn_2005_Nearsightedness of electronic matter.pdf}
}
+@article{pyzer-knappAcceleratingMaterialsDiscovery2022,
+ title = {Accelerating Materials Discovery Using Artificial Intelligence, High Performance Computing and Robotics},
+ author = {Pyzer-Knapp, Edward O. and Pitera, Jed W. and Staar, Peter W. J. and Takeda, Seiji and Laino, Teodoro and Sanders, Daniel P. and Sexton, James and Smith, John R. and Curioni, Alessandro},
+ date = {2022-04-26},
+ journaltitle = {npj Computational Materials},
+ shortjournal = {npj Comput Mater},
+ volume = {8},
+ number = {1},
+ pages = {1--9},
+ publisher = {Nature Publishing Group},
+ issn = {2057-3960},
+ doi = {10.1038/s41524-022-00765-z},
+ url = {https://www.nature.com/articles/s41524-022-00765-z},
+ urldate = {2024-08-01},
+ abstract = {New tools enable new ways of working, and materials science is no exception. In materials discovery, traditional manual, serial, and human-intensive work is being augmented by automated, parallel, and iterative processes driven by Artificial Intelligence (AI), simulation and experimental automation. In this perspective, we describe how these new capabilities enable the acceleration and enrichment of each stage of the discovery cycle. We show, using the example of the development of a novel chemically amplified photoresist, how these technologies’ impacts are amplified when they are used in concert with each other as powerful, heterogeneous workflows.},
+ langid = {english},
+ keywords = {autonomous research systems,for introductions,lab automation,materials acceleration platforms,Materials chemistry,materials discovery,self-driving lab,Theory and computation},
+ file = {/Users/wasmer/Nextcloud/Zotero/Pyzer-Knapp et al_2022_Accelerating materials discovery using artificial intelligence, high.pdf}
+}
+
@online{qamarAtomicClusterExpansion2022,
title = {Atomic Cluster Expansion for Quantum-Accurate Large-Scale Simulations of Carbon},
author = {Qamar, Minaam and Mrovec, Matous and Lysogorskiy, Yury and Bochkarev, Anton and Drautz, Ralf},
date = {2022-10-25},
eprint = {2210.09161},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2210.09161},
url = {http://arxiv.org/abs/2210.09161},
urldate = {2023-01-20},
abstract = {We present an atomic cluster expansion (ACE) for carbon that improves over available classical and machine learning potentials. The ACE is parameterized from an exhaustive set of important carbon structures at extended volume and energy range, computed using density functional theory (DFT). Rigorous validation reveals that ACE predicts accurately a broad range of properties of both crystalline and amorphous carbon phases while being several orders of magnitude more computationally efficient than available machine learning models. We demonstrate the predictive power of ACE on three distinct applications, brittle crack propagation in diamond, evolution of amorphous carbon structures at different densities and quench rates and nucleation and growth of fullerene clusters under high pressure and temperature conditions.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ACE,AML,carbon,DeePMD-kit,descriptors,DFT,GAP,ML,MLP,molecular dynamics},
file = {/Users/wasmer/Nextcloud/Zotero/Qamar et al_2022_Atomic cluster expansion for quantum-accurate large-scale simulations of carbon.pdf;/Users/wasmer/Zotero/storage/SCVIRYIV/2210.html}
}
@@ -13455,7 +14700,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
urldate = {2023-10-06},
isbn = {978-3-031-37195-0 978-3-031-37196-7},
langid = {english},
- keywords = {\_tablet,AML,biomolecules,chemistry,database generation,educational,GNN,learning material,ML,ML-DFA,ML-DFT,ML-ESM,MLP,organic chemistry,review-of-AML,textbook},
+ keywords = {AML,biomolecules,chemistry,database generation,educational,GNN,learning material,ML,ML-DFA,ML-DFT,ML-ESM,MLP,organic chemistry,review-of-AML,textbook},
file = {/Users/wasmer/Nextcloud/Zotero/Qu_Liu_2023_Machine Learning in Molecular Sciences.pdf}
}
@@ -13464,13 +14709,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Rackers, Joshua A. and Tecot, Lucas and Geiger, Mario and Smidt, Tess E.},
date = {2022-02-10},
eprint = {2201.03726},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2201.03726},
url = {http://arxiv.org/abs/2201.03726},
urldate = {2022-07-10},
abstract = {A long-standing goal of science is to accurately solve the Schr\textbackslash "odinger equation for large molecular systems. The poor scaling of current quantum chemistry algorithms on classical computers imposes an effective limit of about a few dozen atoms for which we can calculate molecular electronic structure. We present a machine learning (ML) method to break through this scaling limit and make quantum chemistry calculations of very large systems possible. We show that Euclidean Neural Networks can be trained to predict the electron density with high fidelity from limited data. Learning the electron density allows us to train a machine learning model on small systems and make accurate predictions on large ones. We show that this ML electron density model can break through the quantum scaling limit and calculate the electron density of systems of thousands of atoms with quantum accuracy.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {CCSD(T),charge density,e3nn,ENN,equivariant,GNN,ML,ML-DFT,ML-ESM,molecules,prediction of electron density,script,target: density,transfer learning,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Rackers et al_2022_Cracking the Quantum Scaling Limit with Machine Learned Electron Densities2.pdf;/Users/wasmer/Zotero/storage/NL7QJTKF/2201.html}
}
@@ -13530,13 +14775,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Radhakrishnan, Adityanarayanan and Beaglehole, Daniel and Pandit, Parthe and Belkin, Mikhail},
date = {2023-05-09},
eprint = {2212.13881},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.2212.13881},
url = {http://arxiv.org/abs/2212.13881},
urldate = {2023-09-18},
abstract = {In recent years neural networks have achieved impressive results on many technological and scientific tasks. Yet, the mechanism through which these models automatically select features, or patterns in data, for prediction remains unclear. Identifying such a mechanism is key to advancing performance and interpretability of neural networks and promoting reliable adoption of these models in scientific applications. In this paper, we identify and characterize the mechanism through which deep fully connected neural networks learn features. We posit the Deep Neural Feature Ansatz, which states that neural feature learning occurs by implementing the average gradient outer product to up-weight features strongly related to model output. Our ansatz sheds light on various deep learning phenomena including emergence of spurious features and simplicity biases and how pruning networks can increase performance, the "lottery ticket hypothesis." Moreover, the mechanism identified in our work leads to a backpropagation-free method for feature learning with any machine learning model. To demonstrate the effectiveness of this feature learning mechanism, we use it to enable feature learning in classical, non-feature learning models known as kernel machines and show that the resulting models, which we refer to as Recursive Feature Machines, achieve state-of-the-art performance on tabular data.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Deep learning,General ML,kernel methods,ML,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Radhakrishnan et al_2023_Mechanism of feature learning in deep fully connected networks and kernel.pdf;/Users/wasmer/Zotero/storage/XMIF5REM/2212.html}
}
@@ -13546,13 +14791,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Radhakrishnan, Adityanarayanan and Luyten, Max Ruiz and Prasad, Neha and Uhler, Caroline},
date = {2022-10-31},
eprint = {2211.00227},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2211.00227},
url = {http://arxiv.org/abs/2211.00227},
urldate = {2023-09-18},
abstract = {Transfer learning refers to the process of adapting a model trained on a source task to a target task. While kernel methods are conceptually and computationally simple machine learning models that are competitive on a variety of tasks, it has been unclear how to perform transfer learning for kernel methods. In this work, we propose a transfer learning framework for kernel methods by projecting and translating the source model to the target task. We demonstrate the effectiveness of our framework in applications to image classification and virtual drug screening. In particular, we show that transferring modern kernels trained on large-scale image datasets can result in substantial performance increase as compared to using the same kernel trained directly on the target task. In addition, we show that transfer-learned kernels allow a more accurate prediction of the effect of drugs on cancer cell lines. For both applications, we identify simple scaling laws that characterize the performance of transfer-learned kernels as a function of the number of target examples. We explain this phenomenon in a simplified linear setting, where we are able to derive the exact scaling laws. By providing a simple and effective transfer learning framework for kernel methods, our work enables kernel methods trained on large datasets to be easily adapted to a variety of downstream target tasks.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Computer Science - Machine Learning,Deep learning,General ML,kernel methods,ML,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Radhakrishnan et al_2022_Transfer Learning with Kernel Methods.pdf;/Users/wasmer/Zotero/storage/WGXWDYIS/2211.html}
}
@@ -13563,13 +14808,13 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Raissi, Maziar and Perdikaris, Paris and Karniadakis, George Em},
date = {2017-11-28},
eprint = {1711.10561},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, math, stat},
doi = {10.48550/arXiv.1711.10561},
url = {http://arxiv.org/abs/1711.10561},
urldate = {2023-11-12},
abstract = {We introduce physics informed neural networks -- neural networks that are trained to solve supervised learning tasks while respecting any given law of physics described by general nonlinear partial differential equations. In this two part treatise, we present our developments in the context of solving two main classes of problems: data-driven solution and data-driven discovery of partial differential equations. Depending on the nature and arrangement of the available data, we devise two distinct classes of algorithms, namely continuous time and discrete time models. The resulting neural networks form a new class of data-efficient universal function approximators that naturally encode any underlying physical laws as prior information. In this first part, we demonstrate how these networks can be used to infer solutions to partial differential equations, and obtain physics-informed surrogate models that are fully differentiable with respect to all input coordinates and free parameters.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,CFD,dynamical systems,dynamics,nonlinear dynamics,original publication,PDE,physics-informed ML,PINN,Python,rec-by-bluegel,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Raissi et al_2017_Physics Informed Deep Learning (Part I).pdf;/Users/wasmer/Zotero/storage/8E8HZ9R4/1711.html}
}
@@ -13619,7 +14864,7 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Ramsundar, Bharath and Krishnamurthy, Dilip and Viswanathan, Venkatasubramanian},
date = {2021-09-14},
eprint = {2109.07573},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
publisher = {arXiv},
doi = {10.48550/arXiv.2109.07573},
@@ -13653,14 +14898,14 @@ Subject\_term\_id: magnetic-properties-and-materials},
author = {Reiser, Patrick and Neubert, Marlen and Eberhard, André and Torresi, Luca and Zhou, Chen and Shao, Chen and Metni, Houssam and family=Hoesel, given=Clint, prefix=van, useprefix=true and Schopmans, Henrik and Sommer, Timo and Friederich, Pascal},
date = {2022-08-05},
eprint = {2208.09481},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2208.09481},
url = {http://arxiv.org/abs/2208.09481},
urldate = {2022-09-27},
abstract = {Machine learning plays an increasingly important role in many areas of chemistry and materials science, e.g. to predict materials properties, to accelerate simulations, to design new materials, and to predict synthesis routes of new materials. Graph neural networks (GNNs) are one of the fastest growing classes of machine learning models. They are of particular relevance for chemistry and materials science, as they directly work on a graph or structural representation of molecules and materials and therefore have full access to all relevant information required to characterize materials. In this review article, we provide an overview of the basic principles of GNNs, widely used datasets, and state-of-the-art architectures, followed by a discussion of a wide range of recent applications of GNNs in chemistry and materials science, and concluding with a road-map for the further development and application of GNNs.},
- pubstate = {preprint},
- keywords = {\_tablet,GCN,GNN,molecules,review,solids},
+ pubstate = {prepublished},
+ keywords = {GCN,GNN,molecules,review,solids},
file = {/Users/wasmer/Nextcloud/Zotero/Reiser et al_2022_Graph neural networks for materials science and chemistry.pdf;/Users/wasmer/Zotero/storage/IVEGXDHZ/2208.html}
}
@@ -13683,16 +14928,20 @@ Subject\_term\_id: magnetic-properties-and-materials},
file = {/Users/wasmer/Nextcloud/Zotero/Ren et al_2022_Ligand Optimization of Exchange Interaction in Co(II) Dimer Single Molecule.pdf;/Users/wasmer/Zotero/storage/NZ36VI4U/acs.jpca.html}
}
-@online{riebesellMatbenchDiscoveryCan2023,
- title = {Matbench {{Discovery}} - {{Can}} Machine Learning Identify Stable Crystals?},
- author = {Riebesell, Janosh and Goodall, Rhys E. A. and Jain, Anubhav and Benner, Philipp and Persson, Kristin A. and Lee, Alpha A.},
- date = {2023-06-20},
- url = {https://janosh.github.io/matbench-discovery},
- urldate = {2023-08-29},
- abstract = {We present a new machine learning (ML) evaluation framework for materials stability predictions named Matbench Discovery. Our task closely simulates the deployment of ML energy models in a high-throughput search for stable inorganic crystals. It is accompanied by an interactive leaderboard and a Python package for easy ingestion of our training/test sets into future model submissions. To answer the question which ML methodology performs best at materials discovery, we explore a wide variety of models. Our initial selection ranges from random forests to GNNs, from one-shot predictors to iterative Bayesian optimizers and universal interatomic potentials (UIP) that closely emulate DFT. We find UIPs to be in a class of their own, achieving the highest F1 scores and discovery acceleration factors (DAF) of more than 3, i.e. 3x more stable structures found compared to dummy selection in our already enriched search space. We also identify a sharp disconnect between commonly used regression metrics and more task-relevant classification metrics. CGCNN and MEGNet are worse than dummy regressors, but substantially better than dummy classifiers, suggesting that the field overemphasizes the wrong performance indicators. Our results highlight the need to optimize metrics that measure true stability hit rate improvements and provide valuable insights for maintainers of high throughput materials databases by demonstrating that these models have matured enough to play a vital role as pre-filtering steps to effectively allocate compute budget for DFT relaxations.},
- langid = {english},
- keywords = {AML,Bayesian optimization,benchmark dataset,benchmarking,CGCNN,CHGNet,convex hull,Database,GNN,inorganic materials,library,M3GNet,MatBench,materials,materials project,MEGNet,ML,MLP,platform,todo-tagging,universal potential,voronoi descriptor,with-code,with-data},
- file = {/Users/wasmer/Zotero/storage/Z9GXZ7NV/preprint.html}
+@online{riebesellMatbenchDiscoveryFramework2024,
+ title = {Matbench {{Discovery}} -- {{A}} Framework to Evaluate Machine Learning Crystal Stability Predictions},
+ author = {Riebesell, Janosh and Goodall, Rhys E. A. and Benner, Philipp and Chiang, Yuan and Deng, Bowen and Lee, Alpha A. and Jain, Anubhav and Persson, Kristin A.},
+ date = {2024-02-04},
+ eprint = {2308.14920},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat},
+ doi = {10.48550/arXiv.2308.14920},
+ url = {http://arxiv.org/abs/2308.14920},
+ urldate = {2024-06-16},
+ abstract = {Matbench Discovery simulates the deployment of machine learning (ML) energy models in a high-throughput search for stable inorganic crystals. We address the disconnect between (i) thermodynamic stability and formation energy and (ii) in-domain vs out-of-distribution performance. Alongside this paper, we publish a Python package to aid with future model submissions and a growing online leaderboard with further insights into trade-offs between various performance metrics. To answer the question which ML methodology performs best at materials discovery, our initial release explores a variety of models including random forests, graph neural networks (GNN), one-shot predictors, iterative Bayesian optimizers and universal interatomic potentials (UIP). Ranked best-to-worst by their test set F1 score on thermodynamic stability prediction, we find CHGNet {$>$} M3GNet {$>$} MACE {$>$} ALIGNN {$>$} MEGNet {$>$} CGCNN {$>$} CGCNN+P {$>$} Wrenformer {$>$} BOWSR {$>$} Voronoi tessellation fingerprints with random forest. The top 3 models are UIPs, the winning methodology for ML-guided materials discovery, achieving F1 scores of \textasciitilde 0.6 for crystal stability classification and discovery acceleration factors (DAF) of up to 5x on the first 10k most stable predictions compared to dummy selection from our test set. We also highlight a sharp disconnect between commonly used global regression metrics and more task-relevant classification metrics. Accurate regressors are susceptible to unexpectedly high false-positive rates if those accurate predictions lie close to the decision boundary at 0 eV/atom above the convex hull where most materials are. Our results highlight the need to focus on classification metrics that actually correlate with improved stability hit rate.},
+ pubstate = {prepublished},
+ keywords = {/unread,ALIGNN,AML,benchmarking,CGCNN,CHGNet,community-resource,dataset,GNoME,M3GNet,MACE,materials discovery,materials project,MEGNet,ML,MLP,model comparison,SOTA,universal potential,with-code,with-data,with-demo},
+ file = {/Users/wasmer/Nextcloud/Zotero/Riebesell et al_2024_Matbench Discovery -- A framework to evaluate machine learning crystal.pdf;/Users/wasmer/Zotero/storage/SK6EKID7/2308.html}
}
@inproceedings{riedelEnablingHyperparameterTuningAI2023,
@@ -13719,8 +14968,8 @@ Subject\_term\_id: magnetic-properties-and-materials},
urldate = {2023-07-01},
abstract = {The Atomic Cluster Expansion (ACE) provides a formally complete basis for the local atomic environment. ACE is not limited to representing energies as a function of atomic positions and chemical species, but can be generalized to vectorial or tensorial properties and to incorporate further degrees of freedom (DOF). This is crucial for magnetic materials with potential energy surfaces that depend on atomic positions and atomic magnetic moments simultaneously. In this work, we employ the ACE formalism to develop a non-collinear magnetic ACE parametrization for the prototypical magnetic element Fe. The model is trained on a broad range of collinear and non-collinear magnetic structures calculated using spin density functional theory. We demonstrate that the non-collinear magnetic ACE is able to reproduce not only ground state properties of various magnetic phases of Fe but also the magnetic and lattice excitations that are essential for a correct description of the finite temperature behavior and properties of crystal defects.},
langid = {english},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,AML,Dzyaloshinskii–Moriya interaction,Heisenberg model,higher-order exchange interactions,Jij,linear regression,magnetism,ML,MLP,non-collinear,pacemaker,prediction of Jij,spin-dependent,tensorial target,todo-tagging,with-code},
+ pubstate = {prepublished},
+ keywords = {ACE,AML,Dzyaloshinskii–Moriya interaction,Heisenberg model,higher-order exchange interactions,Jij,linear regression,magnetism,ML,MLP,non-collinear,pacemaker,prediction of Jij,spin-dependent,tensorial target,todo-tagging,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Rinaldi et al_2023_Non-collinear Magnetic Atomic Cluster Expansion for Iron.pdf}
}
@@ -13775,13 +15024,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Roberts, Chase and Milsted, Ashley and Ganahl, Martin and Zalcman, Adam and Fontaine, Bruce and Zou, Yijian and Hidary, Jack and Vidal, Guifre and Leichenauer, Stefan},
date = {2019-05-03},
eprint = {1905.01330},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:hep-th, physics:physics, stat},
doi = {10.48550/arXiv.1905.01330},
url = {http://arxiv.org/abs/1905.01330},
urldate = {2023-06-30},
abstract = {TensorNetwork is an open source library for implementing tensor network algorithms. Tensor networks are sparse data structures originally designed for simulating quantum many-body physics, but are currently also applied in a number of other research areas, including machine learning. We demonstrate the use of the API with applications both physics and machine learning, with details appearing in companion papers.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,General ML,Google,image classification,library,ML,particle physics,Physics ML,prediction of ground-state properties,quantum computing,quantum information,spin,spin chain,spin-dependent,strongly correlated maeterials,tensor network},
file = {/Users/wasmer/Nextcloud/Zotero/Roberts et al_2019_TensorNetwork.pdf;/Users/wasmer/Zotero/storage/PF5AB8AA/1905.html}
}
@@ -13803,7 +15052,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
abstract = {The ability to manipulate electron spin in organic molecular materials offers a new and extremely tantalizing route towards spin electronics, both from fundamental and technological points of view. This is mainly due to the unquestionable advantage of weak spin–orbit and hyperfine interactions in organic molecules, which leads to the possibility of preserving spin-coherence over times and distances much longer than in conventional metals or semiconductors. Here we demonstrate theoretically that organic spin valves, obtained by sandwiching an organic molecule between magnetic contacts, can show a large bias-dependent magnetoresistance and that this can be engineered by an appropriate choice of molecules and anchoring groups. Our results, obtained through a combination of state-of-the-art non-equilibrium transport methods and density functional theory, show that although the magnitude of the effect varies with the details of the molecule, large magnetoresistance can be found both in the tunnelling and the metallic limit.},
issue = {4},
langid = {english},
- keywords = {DFT,magnetism,molecular spintronics,molecules,organic chemistry,physics,review,Spintronics},
+ keywords = {\_tablet,DFT,magnetism,molecular spintronics,molecules,organic chemistry,physics,review,Spintronics},
file = {/Users/wasmer/Nextcloud/Zotero/Rocha et al_2005_Towards molecular spintronics.pdf}
}
@@ -13851,15 +15100,15 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Rossignol, Hugo and Minotakis, Michail and Cobelli, Matteo and Sanvito, Stefano},
date = {2023-08-30},
eprint = {2308.15907},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2308.15907},
url = {http://arxiv.org/abs/2308.15907},
urldate = {2023-09-22},
abstract = {In the search for novel intermetallic ternary alloys, much of the effort goes into performing a large number of ab-initio calculations covering a wide range of compositions and structures. These are essential to build a reliable convex hull diagram. While density functional theory (DFT) provides accurate predictions for many systems, its computational overheads set a throughput limit on the number of hypothetical phases that can be probed. Here, we demonstrate how an ensemble of machine-learning spectral neighbor-analysis potentials (SNAPs) can be integrated into a workflow for the construction of accurate ternary convex hull diagrams, highlighting regions fertile for materials discovery. Our workflow relies on using available binary-alloy data both to train the SNAP models and to create prototypes for ternary phases. From the prototype structures, all unique ternary decorations are created and used to form a pool of candidate compounds. The SNAPs are then used to pre-relax the structures and screen the most favourable prototypes, before using DFT to build the final phase diagram. As constructed, the proposed workflow relies on no extra first-principles data to train the machine-learning surrogate model and yields a DFT-level accurate convex hull. We demonstrate its efficacy by investigating the Cu-Ag-Au and Mo-Ta-W ternary systems.},
- pubstate = {preprint},
- keywords = {/unread,Condensed Matter - Materials Science},
- file = {/Users/wasmer/Zotero/storage/JDTKMCIM/Rossignol et al. - 2023 - Machine-Learning-Assisted Construction of Ternary .pdf;/Users/wasmer/Zotero/storage/2Z9WD4PH/2308.html}
+ pubstate = {prepublished},
+ keywords = {/unread,\_tablet,AML,convex hull,Jacobi-Legendre,ML,phase diagram},
+ file = {/Users/wasmer/Zotero/storage/JDTKMCIM/Rossignol et al_2023_Machine-Learning-Assisted Construction of Ternary Convex Hull Diagrams.pdf;/Users/wasmer/Zotero/storage/2Z9WD4PH/2308.html}
}
@article{rossignolMachineLearningAssistedConstructionTernary2024,
@@ -13877,7 +15126,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
url = {https://doi.org/10.1021/acs.jcim.3c01391},
urldate = {2024-03-31},
abstract = {In the search for novel intermetallic ternary alloys, much of the effort goes into performing a large number of ab initio calculations covering a wide range of compositions and structures. These are essential to building a reliable convex hull diagram. While density functional theory (DFT) provides accurate predictions for many systems, its computational overheads set a throughput limit on the number of hypothetical phases that can be probed. Here, we demonstrate how an ensemble of machine-learning (ML) spectral neighbor-analysis potentials (SNAPs) can be integrated into a workflow for the construction of accurate ternary convex hull diagrams, highlighting regions that are fertile for materials discovery. Our workflow relies on using available binary-alloy data both to train the SNAP models and to create prototypes for ternary phases. From the prototype structures, all unique ternary decorations are created and used to form a pool of candidate compounds. The SNAPs ensemble is then used to prerelax the structures and screen the most favorable prototypes before using DFT to build the final phase diagram. As constructed, the proposed workflow relies on no extra first-principles data to train the ML surrogate model and yields a DFT-level accurate convex hull. We demonstrate its efficacy by investigating the Cu–Ag–Au and Mo–Ta–W ternary systems.},
- keywords = {/unread,AFLOWLIB,alloys,AML,ase,bispectrum,convex hull,ensemble learning,LAMMPS,ML,ML-DFT,ML-ESM,MLP,phase diagram,pymatgen,scikit-learn,SNAP,structure relaxation,surrogate model,ternary systems,with-data},
+ keywords = {/unread,\_tablet,AFLOWLIB,alloys,AML,ase,bispectrum,convex hull,ensemble learning,Jacobi-Legendre,JLCDM,LAMMPS,ML,ML-DFT,ML-ESM,MLP,phase diagram,pymatgen,scikit-learn,SNAP,structure relaxation,surrogate model,ternary systems,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Rossignol et al_2024_Machine-Learning-Assisted Construction of Ternary Convex Hull Diagrams.pdf}
}
@@ -13908,13 +15157,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Rubungo, Andre Niyongabo and Arnold, Craig and Rand, Barry P. and Dieng, Adji Bousso},
date = {2023-10-21},
eprint = {2310.14029},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2310.14029},
url = {http://arxiv.org/abs/2310.14029},
urldate = {2023-12-05},
abstract = {The prediction of crystal properties plays a crucial role in the crystal design process. Current methods for predicting crystal properties focus on modeling crystal structures using graph neural networks (GNNs). Although GNNs are powerful, accurately modeling the complex interactions between atoms and molecules within a crystal remains a challenge. Surprisingly, predicting crystal properties from crystal text descriptions is understudied, despite the rich information and expressiveness that text data offer. One of the main reasons is the lack of publicly available data for this task. In this paper, we develop and make public a benchmark dataset (called TextEdge) that contains text descriptions of crystal structures with their properties. We then propose LLM-Prop, a method that leverages the general-purpose learning capabilities of large language models (LLMs) to predict the physical and electronic properties of crystals from their text descriptions. LLM-Prop outperforms the current state-of-the-art GNN-based crystal property predictor by about 4\% in predicting band gap, 3\% in classifying whether the band gap is direct or indirect, and 66\% in predicting unit cell volume. LLM-Prop also outperforms a finetuned MatBERT, a domain-specific pre-trained BERT model, despite having 3 times fewer parameters. Our empirical results may highlight the current inability of GNNs to capture information pertaining to space group symmetry and Wyckoff sites for accurate crystal property prediction.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AML,language models,library,LLM,materials,ML,todo-tagging,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Rubungo et al_2023_LLM-Prop.pdf;/Users/wasmer/Zotero/storage/BLZZVYKN/2310.html}
}
@@ -13924,13 +15173,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Ruhe, David and Gupta, Jayesh K. and family=Keninck, given=Steven, prefix=de, useprefix=true and Welling, Max and Brandstetter, Johannes},
date = {2023-05-29},
eprint = {2302.06594},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2302.06594},
url = {http://arxiv.org/abs/2302.06594},
urldate = {2023-08-22},
abstract = {We propose Geometric Clifford Algebra Networks (GCANs) for modeling dynamical systems. GCANs are based on symmetry group transformations using geometric (Clifford) algebras. We first review the quintessence of modern (plane-based) geometric algebra, which builds on isometries encoded as elements of the \$\textbackslash mathrm\{Pin\}(p,q,r)\$ group. We then propose the concept of group action layers, which linearly combine object transformations using pre-specified group actions. Together with a new activation and normalization scheme, these layers serve as adjustable \$\textbackslash textit\{geometric templates\}\$ that can be refined via gradient descent. Theoretical advantages are strongly reflected in the modeling of three-dimensional rigid body transformations as well as large-scale fluid dynamics simulations, showing significantly improved performance over traditional methods.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Ruhe et al_2023_Geometric Clifford Algebra Networks.pdf;/Users/wasmer/Zotero/storage/66MRFXSJ/2302.html}
}
@@ -13967,7 +15216,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
url = {https://link.aps.org/doi/10.1103/PhysRevLett.108.058301},
urldate = {2021-07-10},
abstract = {We introduce a machine learning model to predict atomization energies of a diverse set of organic molecules, based on nuclear charges and atomic positions only. The problem of solving the molecular Schrödinger equation is mapped onto a nonlinear statistical regression problem of reduced complexity. Regression models are trained on and compared to atomization energies computed with hybrid density-functional theory. Cross validation over more than seven thousand organic molecules yields a mean absolute error of ∼10 kcal/mol. Applicability is demonstrated for the prediction of molecular atomization potential energy curves.},
- keywords = {Coulomb matrix,descriptors,ML,original publication},
+ keywords = {AML,Coulomb matrix,descriptors,ML,original publication},
file = {/Users/wasmer/Nextcloud/Zotero/Rupp et al_2012_Fast and Accurate Modeling of Molecular Atomization Energies with Machine.pdf;/Users/wasmer/Zotero/storage/AP7Y6JEW/PhysRevLett.108.html}
}
@@ -13985,7 +15234,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
urldate = {2022-05-13},
abstract = {Models that combine quantum mechanics (QM) with machine learning (ML) promise to deliver the accuracy of QM at the speed of ML. This hands-on tutorial introduces the reader to QM/ML models based on kernel learning, an elegant, systematically nonlinear form of ML. Pseudocode and a reference implementation are provided, enabling the reader to reproduce results from recent publications where atomization energies of small organic molecules are predicted using kernel ridge regression. © 2015 Wiley Periodicals, Inc.},
langid = {english},
- keywords = {\_tablet,Coulomb matrix,GPR,kernel methods,KRR,ML,models,tutorial},
+ keywords = {Coulomb matrix,GPR,kernel methods,KRR,ML,models,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Rupp_2015_Machine learning for quantum mechanics in a nutshell.pdf;/Users/wasmer/Zotero/storage/7CP5YBAD/qua.html}
}
@@ -14039,18 +15288,36 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
file = {/Users/wasmer/Nextcloud/Zotero/Rüßmann et al_2022_The AiiDA-Spirit Plugin for Automated Spin-Dynamics Simulations and Multi-Scale.pdf}
}
+@article{russmannDensityFunctionalBogoliubovde2022,
+ title = {Density Functional {{Bogoliubov-de Gennes}} Analysis of Superconducting {{Nb}} and {{Nb}}(110) Surfaces},
+ author = {Rüßmann, Philipp and Blügel, Stefan},
+ date = {2022-03-31},
+ journaltitle = {Physical Review B},
+ shortjournal = {Phys. Rev. B},
+ volume = {105},
+ number = {12},
+ pages = {125143},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevB.105.125143},
+ url = {https://link.aps.org/doi/10.1103/PhysRevB.105.125143},
+ urldate = {2024-08-03},
+ abstract = {We report on the implementation of the Bogoliubov-de Gennes method into the JuKKR code, an implementation of the relativistic all-electron, full-potential Korringa-Kohn-Rostoker Green function method, which allows a material-specific description of inhomogeneous superconductors and heterostructures on the basis of density functional theory. We describe the formalism and report on calculations for the ð‘ -wave superconductor Nb, a potential component of the materials platform enabling the realization of the Majorana zero modes in the field of topological quantum computing. We compare the properties of the superconducting state both in the bulk and for (110) surfaces. We compare slab calculations for different thicknesses, the effect of surface relaxations, and the influence of a softening of phonon modes on the surface for the resulting superconducting gap.},
+ keywords = {/unread,BdG,BdG-DFT,DFT,FZJ,impurity embedding,JuKKR,KKR,PGI,PGI-1/IAS-1,superconductor,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Rüßmann_Blügel_2022_Density functional Bogoliubov-de Gennes analysis of superconducting Nb and.pdf;/Users/wasmer/Zotero/storage/IINPIHFD/PhysRevB.105.html}
+}
+
@online{russmannDensityfunctionalDescriptionMaterials2023,
title = {Density-Functional Description of Materials for Topological Qubits and Superconducting Spintronics},
author = {Rüßmann, Philipp and Silva, David Antognini and Hemmati, Mohammad and Klepetsanis, Ilias and Trauzettel, Björn and Mavropoulos, Phivos and Blügel, Stefan},
date = {2023-08-14},
eprint = {2308.07383},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2308.07383},
url = {http://arxiv.org/abs/2308.07383},
urldate = {2023-08-19},
abstract = {Interfacing superconductors with magnetic or topological materials offers a playground where novel phenomena like topological superconductivity, Majorana zero modes, or superconducting spintronics are emerging. In this work, we discuss recent developments in the Kohn-Sham Bogoliubov-de Gennes method, which allows to perform material-specific simulations of complex superconducting heterostructures on the basis of density functional theory. As a model system we study magnetically-doped Pb. In our analysis we focus on the interplay of magnetism and superconductivity. This combination leads to Yu-Shiba-Rusinov (YSR) in-gap bound states at magnetic defects and the breakdown of superconductivity at larger impurity concentrations. Moreover, the influence of spin-orbit coupling and on orbital splitting of YSR states as well as the appearance of a triplet component in the order parameter is discussed. These effects can be exploited in S/F/S-type devices (S=superconductor, F=ferromagnet) in the field of superconducting spintronics.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {BdG,CPA,defects,DFT,FZJ,impurity embedding,JuKKR,KKR,KS-BdG,MZM,PGI,PGI-1/IAS-1,physics,quantum materials,SOC,spintronics,superconducting spitronics,superconductor,Topological Superconductor,Yu-Shiba-Rusinov},
file = {/Users/wasmer/Nextcloud/Zotero/Rüßmann et al_2023_Density-functional description of materials for topological qubits and.pdf;/Users/wasmer/Zotero/storage/5NLHWP9G/2308.html}
}
@@ -14079,7 +15346,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
urldate = {2022-05-30},
abstract = {The quasiparticle interference (QPI) technique is a powerful tool that allows to uncover the structure and properties of electronic structure of a material combined with scattering properties of defects at surfaces. Recently, this technique has been pivotal in proving the unique properties of the surface state of topological insulators which manifests itself in the absence of backscattering. Herein, a Green function-based formalism is derived for the ab initio computation of Fourier-transformed QPI images. The efficiency of the new implementation is shown at the examples of QPI that forms around magnetic and nonmagnetic defects at the Bi2Te3 surface. This method allows a deepened understanding of the scattering properties of topologically protected electrons off defects and is a useful tool in the study of quantum materials in the future.},
langid = {english},
- keywords = {\_tablet,density functional theory,impurity scattering,Korringa–Kohn–Rostoker,quasiparticle interferences,topological insulators},
+ keywords = {density functional theory,impurity scattering,Korringa–Kohn–Rostoker,quasiparticle interferences,topological insulators},
file = {/Users/wasmer/Nextcloud/Zotero/Rüßmann et al_2021_Ab Initio Theory of Fourier-Transformed Quasiparticle Interference Maps and.pdf}
}
@@ -14088,13 +15355,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Rüßmann, Philipp and Bahari, Masoud and Blügel, Stefan and Trauzettel, Björn},
date = {2023-07-26},
eprint = {2307.13990},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2307.13990},
url = {http://arxiv.org/abs/2307.13990},
urldate = {2023-08-19},
abstract = {Multi-band effects in hybrid structures provide a rich playground for unconventional superconductivity. We combine two complementary approaches based on density-functional theory (DFT) and effective low-energy model theory in order to investigate the proximity effect in a Rashba surface state in contact to an \$s\$-wave superconductor. We discuss these synergistic approaches and combine the effective model and DFT analysis at the example of a Au/Al heterostructure. This allows us to predict finite-energy superconducting pairing due to the interplay of the Rashba surface state of Au, and hybridization with the electronic structure of superconducting Al. We investigate the nature of the induced superconducting pairing and quantify its mixed singlet-triplet character. Our findings demonstrate general recipes to explore real material systems that exhibit inter-orbital pairing away from the Fermi energy.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,BdG,DFT,FZJ,heterostructures,juKKR,KKR,Mat4QIT,mesoscopic,MZM,PGI,PGI-1/IAS-1,physics,quantum materials,Rashba effect,S-wave Superconductors,SOC,superconductor},
file = {/Users/wasmer/Nextcloud/Zotero/Rüßmann et al_2023_Inter-orbital Cooper pairing at finite energies in Rashba surface states.pdf;/Users/wasmer/Zotero/storage/THRDSW2F/2307.html}
}
@@ -14104,13 +15371,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Rüßmann, Philipp and Blügel, Stefan},
date = {2022-08-30},
eprint = {2208.14289},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2208.14289},
url = {http://arxiv.org/abs/2208.14289},
urldate = {2023-08-19},
abstract = {Interfacing a topological insulator (TI) with an \$s\$-wave superconductor (SC) is a promising material platform that offers the possibility to realize a topological superconductor through which Majorana-based topologically protected qubits can be engineered. In our computational study of the prototypical SC/TI interface between Nb and Bi\$\_2\$Te\$\_3\$, we identify the benefits and possible bottlenecks of this potential Majorana material platform. Bringing Nb in contact with the TI film induces charge doping from the SC to the TI, which shifts the Fermi level into the TI conduction band. For thick TI films, this results in band bending leading to the population of trivial TI quantum-well states at the interface. In the superconducting state, we uncover that the topological surface state experiences a sizable superconducting gap-opening at the SC/TI interface, which is furthermore robust against fluctuations of the Fermi energy. We also show that the trivial interface state is only marginally proximitized, potentially obstructing the realization of Majorana-based qubits in this material platform.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {PGI-1/IAS-1,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Rüßmann_Blügel_2022_Proximity induced superconductivity in a topological insulator.pdf;/Users/wasmer/Zotero/storage/5Q45YH6R/2208.html}
}
@@ -14126,8 +15393,8 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
abstract = {This thesis provides a detailed microscopic understanding of the impurity scattering of topologically protected electrons, which are studied within the example of strong threedimensional topological insulators (Tls) and type-II Weyl semimetals. The immense research interest in the recent past in topological materials is to a large extend due to the fact that their unconventional electronic surface states are robust against perturbations, such as surface structural relaxations or defects. One of the most intringuing physical properties of topological surface states in Tls is the forbidden backscattering off time-reversal invariant defects, which makes Tl materials very promising candidates for future low-power electronics or quantum information technology. [...] Rüßmann, Philipp},
isbn = {9783958063365},
langid = {english},
- keywords = {juKKR,KKR,PGI-1/IAS-1,thesis},
- file = {/Users/wasmer/Nextcloud/Zotero/Rüßmann_2018_Spin scattering of topologically protected electrons at defects4.pdf;/Users/wasmer/Zotero/storage/T7V45S9S/850306.html}
+ keywords = {\_tablet,juKKR,KKR,PGI-1/IAS-1,thesis},
+ file = {/Users/wasmer/Nextcloud/Zotero/Rüßmann_2018_Spin scattering of topologically protected electrons at defects.pdf;/Users/wasmer/Zotero/storage/T7V45S9S/850306.html}
}
@article{ryczkoDeepLearningDensityfunctional2019,
@@ -14223,13 +15490,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Salzbrenner, Pascal T. and Joo, Se Hun and Conway, Lewis J. and Cooke, Peter I. C. and Zhu, Bonan and Matraszek, Milosz P. and Witt, William C. and Pickard, Chris J.},
date = {2023-06-10},
eprint = {2306.06475},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2306.06475},
url = {http://arxiv.org/abs/2306.06475},
urldate = {2023-06-26},
abstract = {Machine-learned interatomic potentials are fast becoming an indispensable tool in computational materials science. One approach is the ephemeral data-derived potential (EDDP), which was designed to accelerate atomistic structure prediction. The EDDP is simple and cost-efficient. It relies on training data generated in small unit cells and is fit using a lightweight neural network, leading to smooth interactions which exhibit the robust transferability essential for structure prediction. Here, we present a variety of applications of EDDPs, enabled by recent developments of the open-source EDDP software. New features include interfaces to phonon and molecular dynamics codes, as well as deployment of the ensemble deviation for estimating the confidence in EDDP predictions. Through case studies ranging from elemental carbon and lead to the binary scandium hydride and the ternary zinc cyanide, we demonstrate that EDDPs can be trained to cover wide ranges of pressures and stoichiometries, and used to evaluate phonons, phase diagrams, superionicity, and thermal expansion. These developments complement continued success in accelerated structure prediction.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,CASTEP,DFT,ensemble learning,HDNNP,Julia,Lennard-Jones,library,materials,ML,MLP,prediction of energy,PW,structure relaxation,with-code,workflows},
file = {/Users/wasmer/Nextcloud/Zotero/Salzbrenner et al_2023_Developments and Further Applications of Ephemeral Data Derived Potentials.pdf;/Users/wasmer/Zotero/storage/FAY5AQLM/2306.html}
}
@@ -14261,7 +15528,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Samuel, Sheeba and Löffler, Frank and König-Ries, Birgitta},
date = {2020-06-22},
eprint = {2006.12117},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
url = {http://arxiv.org/abs/2006.12117},
urldate = {2021-10-21},
@@ -14294,13 +15561,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Santos, Kylee and Moore, Stan and Oppelstrup, Tomas and Sharifian, Amirali and Sharapov, Ilya and Thompson, Aidan and Kalchev, Delyan Z. and Perez, Danny and Schreiber, Robert and Pakin, Scott and Leon, Edgar A. and Laros III, James H. and James, Michael and Rajamanickam, Sivasankaran},
date = {2024-05-13},
eprint = {2405.07898},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2405.07898},
url = {http://arxiv.org/abs/2405.07898},
urldate = {2024-05-16},
abstract = {Molecular dynamics (MD) simulations have transformed our understanding of the nanoscale, driving breakthroughs in materials science, computational chemistry, and several other fields, including biophysics and drug design. Even on exascale supercomputers, however, runtimes are excessive for systems and timescales of scientific interest. Here, we demonstrate strong scaling of MD simulations on the Cerebras Wafer-Scale Engine. By dedicating a processor core for each simulated atom, we demonstrate a 179-fold improvement in timesteps per second versus the Frontier GPU-based Exascale platform, along with a large improvement in timesteps per unit energy. Reducing every year of runtime to two days unlocks currently inaccessible timescales of slow microstructure transformation processes that are critical for understanding material behavior and function. Our dataflow algorithm runs Embedded Atom Method (EAM) simulations at rates over 270,000 timesteps per second for problems with up to 800k atoms. This demonstrated performance is unprecedented for general-purpose processing cores.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AI accelerators,alternative approaches,alternative to ML,Cerebras,computational cost,computational intensity,EAM,exascale,HPC,MD,scaling,speedup,strong scaling,wafer-scale integration},
file = {/Users/wasmer/Nextcloud/Zotero/Santos et al_2024_Breaking the Molecular Dynamics Timescale Barrier Using a Wafer-Scale System.pdf;/Users/wasmer/Zotero/storage/F3F5SE8S/2405.html}
}
@@ -14321,7 +15588,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
urldate = {2023-06-30},
abstract = {The electron spin made its debut in the device world only two decades ago but today our ability of detecting the spin state of a moving electron underpins the entire magnetic data storage industry. This technological revolution has been driven by a constant improvement in our understanding on how spins can be injected, manipulated and detected in the solid state, a field which is collectively named Spintronics. Recently a number of pioneering experiments and theoretical works suggest that organic materials can offer similar and perhaps superior performances in making spin-devices than the more conventional inorganic metals and semiconductors. Furthermore they can pave the way for radically new device concepts. This is Molecular Spintronics, a blossoming research area aimed at exploring how the unique properties of the organic world can marry the requirements of spin-devices. Importantly, after a first phase, where most of the research was focussed on exporting the concepts of inorganic spintronics to organic materials, the field has moved to a more mature age, where the exploitation of the unique properties of molecules has begun to emerge. Molecular spintronics now collects a diverse and interdisciplinary community ranging from device physicists to synthetic chemists to surface scientists. In this critical review, I will survey this fascinating, rapidly evolving, field with a particular eye on new directions and opportunities. The main differences and challenges with respect to standard spintronics will be discussed and so will be the potential cross-fertilization with other fields (177 references).},
langid = {english},
- keywords = {magnetism,molecular spintronics,molecules,organic chemistry,review,spintronics},
+ keywords = {\_tablet,magnetism,molecular spintronics,molecules,organic chemistry,review,spintronics},
file = {/Users/wasmer/Nextcloud/Zotero/Sanvito_2011_Molecular spintronics.pdf}
}
@@ -14330,13 +15597,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Sasse, Leonard and Nicolaisen-Sobesky, Eliana and Dukart, Juergen and Eickhoff, Simon B. and Götz, Michael and Hamdan, Sami and Komeyer, Vera and Kulkarni, Abhijit and Lahnakoski, Juha and Love, Bradley C. and Raimondo, Federico and Patil, Kaustubh R.},
date = {2023-11-07},
eprint = {2311.04179},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2311.04179},
url = {http://arxiv.org/abs/2311.04179},
urldate = {2024-01-14},
abstract = {Machine learning (ML) provides powerful tools for predictive modeling. ML's popularity stems from the promise of sample-level prediction with applications across a variety of fields from physics and marketing to healthcare. However, if not properly implemented and evaluated, ML pipelines may contain leakage typically resulting in overoptimistic performance estimates and failure to generalize to new data. This can have severe negative financial and societal implications. Our aim is to expand understanding associated with causes leading to leakage when designing, implementing, and evaluating ML pipelines. Illustrated by concrete examples, we provide a comprehensive overview and discussion of various types of leakage that may arise in ML pipelines.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,best practices,data leakage,FZJ,General ML,ML,MLOps,workflows},
file = {/Users/wasmer/Nextcloud/Zotero/Sasse et al_2023_On Leakage in Machine Learning Pipelines.pdf;/Users/wasmer/Zotero/storage/VLTE5EB5/2311.html}
}
@@ -14363,13 +15630,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Satorras, Victor Garcia and Hoogeboom, Emiel and Welling, Max},
date = {2022-02-16},
eprint = {2102.09844},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.2102.09844},
url = {http://arxiv.org/abs/2102.09844},
urldate = {2023-08-22},
abstract = {This paper introduces a new model to learn graph neural networks equivariant to rotations, translations, reflections and permutations called E(n)-Equivariant Graph Neural Networks (EGNNs). In contrast with existing methods, our work does not require computationally expensive higher-order representations in intermediate layers while it still achieves competitive or better performance. In addition, whereas existing methods are limited to equivariance on 3 dimensional spaces, our model is easily scaled to higher-dimensional spaces. We demonstrate the effectiveness of our method on dynamical systems modelling, representation learning in graph autoencoders and predicting molecular properties.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Satorras et al_2022_E(n) Equivariant Graph Neural Networks.pdf;/Users/wasmer/Zotero/storage/TKSFFYBM/2102.html}
}
@@ -14379,13 +15646,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Satorras, Victor Garcia and Hoogeboom, Emiel and Fuchs, Fabian B. and Posner, Ingmar and Welling, Max},
date = {2022-01-14},
eprint = {2105.09016},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2105.09016},
url = {http://arxiv.org/abs/2105.09016},
urldate = {2023-08-22},
abstract = {This paper introduces a generative model equivariant to Euclidean symmetries: E(n) Equivariant Normalizing Flows (E-NFs). To construct E-NFs, we take the discriminative E(n) graph neural networks and integrate them as a differential equation to obtain an invertible equivariant function: a continuous-time normalizing flow. We demonstrate that E-NFs considerably outperform baselines and existing methods from the literature on particle systems such as DW4 and LJ13, and on molecules from QM9 in terms of log-likelihood. To the best of our knowledge, this is the first flow that jointly generates molecule features and positions in 3D.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Satorras et al_2022_E(n) Equivariant Normalizing Flows.pdf;/Users/wasmer/Zotero/storage/8GNMV24K/2105.html}
}
@@ -14417,7 +15684,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Sauceda, Huziel E. and Gálvez-González, Luis E. and Chmiela, Stefan and Paz-Borbón, Lauro Oliver and Müller, Klaus-Robert and Tkatchenko, Alexandre},
date = {2021-06-08},
eprint = {2106.04229},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, physics:quant-ph},
url = {http://arxiv.org/abs/2106.04229},
urldate = {2021-06-17},
@@ -14432,14 +15699,14 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Schaaf, Lars and Fako, Edvin and De, Sandip and Schäfer, Ansgar and Csányi, Gábor},
date = {2023-01-24},
eprint = {2301.09931},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2301.09931},
url = {http://arxiv.org/abs/2301.09931},
urldate = {2023-04-13},
abstract = {In this study, we introduce an automatic training protocol for developing machine learning force fields (MLFFs) that can accurately determine reaction barriers for a given catalytic reaction pathway. The protocol is demonstrated through its application to the eleven-step hydrogenation of carbon dioxide to methanol over an indium oxide catalyst. The training set is iteratively expanded with active learning, using the model's uncertainty estimates to sample novel configurations that are chemically relevant. Our final force field obtains reaction barriers that are within 0.05 eV of those obtained through Density Functional Theory (DFT) calculations. Additionally, we examine two extrapolation tasks. Firstly, we demonstrate that with only a few extra single-point DFT calculations, we can accurately capture the adsorption energy for all eleven reaction intermediates on platinum-doped surfaces. Secondly, we show that MLFFs can be used to identify a wide range of low-energy adsorption configurations that are thermodynamically relevant. This abundance of adsorbate geometries highlights the need for fast and accurate alternatives to direct ab-initio simulations.},
- pubstate = {preprint},
- keywords = {\_tablet,active learning,active learning protocol,AML,chemistry,GAP,Gaussian process,GPR,iterative learning scheme,MACE,ML,ML-FF,MLP,SOAP},
+ pubstate = {prepublished},
+ keywords = {active learning,active learning protocol,AML,chemistry,GAP,Gaussian process,GPR,iterative learning scheme,MACE,ML,ML-FF,MLP,SOAP},
file = {/Users/wasmer/Nextcloud/Zotero/Schaaf et al_2023_Accurate Reaction Barriers for Catalytic Pathways.pdf;/Users/wasmer/Zotero/storage/LACJS8K5/2301.html}
}
@@ -14448,13 +15715,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Schaarschmidt, Michael and Riviere, Morgane and Ganose, Alex M. and Spencer, James S. and Gaunt, Alexander L. and Kirkpatrick, James and Axelrod, Simon and Battaglia, Peter W. and Godwin, Jonathan},
date = {2022-09-26},
eprint = {2209.12466},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2209.12466},
url = {http://arxiv.org/abs/2209.12466},
urldate = {2023-04-03},
abstract = {We present evidence that learned density functional theory (``DFT'') force fields are ready for ground state catalyst discovery. Our key finding is that relaxation using forces from a learned potential yields structures with similar or lower energy to those relaxed using the RPBE functional in over 50\textbackslash\% of evaluated systems, despite the fact that the predicted forces differ significantly from the ground truth. This has the surprising implication that learned potentials may be ready for replacing DFT in challenging catalytic systems such as those found in the Open Catalyst 2020 dataset. Furthermore, we show that a force field trained on a locally harmonic energy surface with the same minima as a target DFT energy is also able to find lower or similar energy structures in over 50\textbackslash\% of cases. This ``Easy Potential'' converges in fewer steps than a standard model trained on true energies and forces, which further accelerates calculations. Its success illustrates a key point: learned potentials can locate energy minima even when the model has high force errors. The main requirement for structure optimisation is simply that the learned potential has the correct minima. Since learned potentials are fast and scale linearly with system size, our results open the possibility of quickly finding ground states for large systems.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,Database,DeepMind,Easy Potential,GNN,Graph Net Simulator,Jax,ML,ML-DFT,ML-FF,MLP,MPNN,OC20,Open Catalyst,original publication,PES,structure relaxation},
file = {/Users/wasmer/Nextcloud/Zotero/Schaarschmidt et al_2022_Learned Force Fields Are Ready For Ground State Catalyst Discovery.pdf;/Users/wasmer/Zotero/storage/8QZN3D56/2209.html}
}
@@ -14476,7 +15743,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
abstract = {We employ machine learning to derive tight-binding parametrizations for the electronic structure of defects. We test several machine learning methods that map the atomic and electronic structure of a defect onto a sparse tight-binding parameterization. Since Multi-layer perceptrons (i.e., feed-forward neural networks) perform best we adopt them for our further investigations. We demonstrate the accuracy of our parameterizations for a range of important electronic structure properties such as band structure, local density of states, transport and level spacing simulations for two common defects in single layer graphene. Our machine learning approach achieves results comparable to maximally localized Wannier functions (i.e., DFT accuracy) without prior knowledge about the electronic structure of the defects while also allowing for a reduced interaction range which substantially reduces calculation time. It is general and can be applied to a wide range of other materials, enabling accurate large-scale simulations of material properties in the presence of different defects.},
issue = {1},
langid = {english},
- keywords = {/unread,Electronic properties and devices,Electronic properties and materials,Electronic structure},
+ keywords = {AML,defects,DNN,graphene,inverse problem,ML,point defects,prediction from energy,prediction of Hamiltonian matrix,supercell,TB,tight binding,transport properties,transport simulation,vacancies,Wannier},
file = {/Users/wasmer/Nextcloud/Zotero/Schattauer et al_2022_Machine learning sparse tight-binding parameters for defects.pdf}
}
@@ -14485,13 +15752,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Scherbela, Michael and Gerard, Leon and Grohs, Philipp},
date = {2023-03-17},
eprint = {2303.09949},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2303.09949},
url = {http://arxiv.org/abs/2303.09949},
urldate = {2023-04-17},
abstract = {Deep neural networks have become a highly accurate and powerful wavefunction ansatz in combination with variational Monte Carlo methods for solving the electronic Schr\textbackslash "odinger equation. However, despite their success and favorable scaling, these methods are still computationally too costly for wide adoption. A significant obstacle is the requirement to optimize the wavefunction from scratch for each new system, thus requiring long optimization. In this work, we propose a novel neural network ansatz, which effectively maps uncorrelated, computationally cheap Hartree-Fock orbitals, to correlated, high-accuracy neural network orbitals. This ansatz is inherently capable of learning a single wavefunction across multiple compounds and geometries, as we demonstrate by successfully transferring a wavefunction model pre-trained on smaller fragments to larger compounds. Furthermore, we provide ample experimental evidence to support the idea that extensive pre-training of a such a generalized wavefunction model across different compounds and geometries could lead to a foundation wavefunction model. Such a model could yield high-accuracy ab-initio energies using only minimal computational effort for fine-tuning and evaluation of observables.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,chemical species scaling problem,DeepErwin,equivariant,FermiNet,few-shot learning,foundation models,HFT,Many-body theory,ML,ML-ESM,ML-QM,ML-QMBP,molecules,MPNN,multi-species,PauliNet,PES,prediction of wavefunction,pretrained models,transfer learning,VMC},
file = {/Users/wasmer/Nextcloud/Zotero/Scherbela et al_2023_Towards a Foundation Model for Neural Network Wavefunctions.pdf;/Users/wasmer/Zotero/storage/YVXYKDB3/2303.html}
}
@@ -14501,7 +15768,7 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Scherbela, Michael and Reisenhofer, Rafael and Gerard, Leon and Marquetand, Philipp and Grohs, Philipp},
date = {2021-12-17},
eprint = {2105.08351},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
url = {http://arxiv.org/abs/2105.08351},
urldate = {2022-03-28},
@@ -14551,6 +15818,24 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
file = {/Users/wasmer/Nextcloud/Zotero/Schlenz_Sandfeld_2022_Applications of Machine Learning to the Study of Crystalline Materials.pdf}
}
+@article{schloglRoleChemistryEnergy2010,
+ title = {The {{Role}} of {{Chemistry}} in the {{Energy Challenge}}},
+ author = {Schlögl, Robert},
+ date = {2010},
+ journaltitle = {ChemSusChem},
+ volume = {3},
+ number = {2},
+ pages = {209--222},
+ issn = {1864-564X},
+ doi = {10.1002/cssc.200900183},
+ url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/cssc.200900183},
+ urldate = {2024-08-01},
+ abstract = {Chemistry with its key targets of providing materials and processes for conversion of matter is at the center stage of the energy challenge. Most energy conversion systems work on (bio)chemical energy carriers and require for their use suitable process and material solutions. The enormous scale of their application demands optimization beyond the incremental improvement of empirical discoveries. Knowledge-based systematic approaches are mandatory to arrive at scalable and sustainable solutions. Chemistry for energy, “ENERCHEM†contributes in many ways already today to the use of fossil energy carriers. Optimization of these processes exemplified by catalysis for fuels and chemicals production or by solid-state lightning can contribute in the near future substantially to the dual challenge of energy use and climate protection being in fact two sides of the same challenge. The paper focuses on the even greater role that ENERCHEM will have to play in the era of renewable energy systems where the storage of solar energy in chemical carries and batteries is a key requirement. A multidisciplinary and diversified approach is suggested to arrive at a stable and sustainable system of energy conversion processes. The timescales for transformation of the present energy scenario will be decades and the resources will be of global economic dimensions. ENERCHEM will have to provide the reliable basis for such technologies based on deep functional understanding.},
+ langid = {english},
+ keywords = {chemistry,energy challenge,energy materials,for introductions},
+ file = {/Users/wasmer/Nextcloud/Zotero/Schlögl_2010_The Role of Chemistry in the Energy Challenge.pdf;/Users/wasmer/Zotero/storage/UCQIWIUH/cssc.html}
+}
+
@article{schmidhuberDeepLearningNeural2015,
title = {Deep Learning in Neural Networks: {{An}} Overview},
shorttitle = {Deep Learning in Neural Networks},
@@ -14591,13 +15876,13 @@ Subject\_term\_id: condensed-matter-physics;history;quantum-physics},
author = {Schmidt, Jonathan and Hoffmann, Noah and Wang, Hai-Chen and Borlido, Pedro and Carriço, Pedro J. M. A. and Cerqueira, Tiago F. T. and Botti, Silvana and Marques, Miguel A. L.},
date = {2022-10-02},
eprint = {2210.00579},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2210.00579},
url = {http://arxiv.org/abs/2210.00579},
urldate = {2023-04-04},
abstract = {Crystal-graph attention networks have emerged recently as remarkable tools for the prediction of thermodynamic stability and materials properties from unrelaxed crystal structures. Previous networks trained on two million materials exhibited, however, strong biases originating from underrepresented chemical elements and structural prototypes in the available data. We tackled this issue computing additional data to provide better balance across both chemical and crystal-symmetry space. Crystal-graph networks trained with this new data show unprecedented generalization accuracy, and allow for reliable, accelerated exploration of the whole space of inorganic compounds. We applied this universal network to perform machine-learning assisted high-throughput materials searches including 2500 binary and ternary structure prototypes and spanning about 1 billion compounds. After validation using density-functional theory, we uncover in total 19512 additional materials on the convex hull of thermodynamic stability and \textasciitilde 150000 compounds with a distance of less than 50 meV/atom from the hull. Combining again machine learning and ab-initio methods, we finally evaluate the discovered materials for applications as superconductors, superhard materials, and we look for candidates with large gap deformation potentials, finding several compounds with extreme values of these properties.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,atomate,CGAT,convex hull,crystal graph,crystal structure,crystal structure prediction,database generation,GAT,GATN,HTC,materials discovery,ML,MPNN,PBE,perovskites,polymorphs,prediction of Curie temperature,prediction of structure,superconductor,ternary systems,thermodynamic stability,transfer learning,VASP},
file = {/Users/wasmer/Nextcloud/Zotero/Schmidt et al_2022_Large-scale machine-learning-assisted exploration of the whole materials space.pdf;/Users/wasmer/Zotero/storage/Z7LCYKCS/2210.html}
}
@@ -14679,7 +15964,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
url = {https://doi.org/10.1021/acs.nanolett.1c04055},
urldate = {2022-04-16},
abstract = {The integration of semiconductor Josephson junctions (JJs) in superconducting quantum circuits provides a versatile platform for hybrid qubits and offers a powerful way to probe exotic quasiparticle excitations. Recent proposals for using circuit quantum electrodynamics (cQED) to detect topological superconductivity motivate the integration of novel topological materials in such circuits. Here, we report on the realization of superconducting transmon qubits implemented with (Bi0.06Sb0.94)2Te3 topological insulator (TI) JJs using ultrahigh vacuum fabrication techniques. Microwave losses on our substrates, which host monolithically integrated hardmasks used for the selective area growth of TI nanostructures, imply microsecond limits to relaxation times and, thus, their compatibility with strong-coupling cQED. We use the cavity–qubit interaction to show that the Josephson energy of TI-based transmons scales with their JJ dimensions and demonstrate qubit control as well as temporal quantum coherence. Our results pave the way for advanced investigations of topological materials in both novel Josephson and topological qubits.},
- keywords = {\_tablet,experimental,FZJ,PGI,PGI-9,quantum computing,superconductor,topological insulator},
+ keywords = {experimental,FZJ,PGI,PGI-9,quantum computing,superconductor,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Schmitt et al_2022_Integration of Topological Insulator Josephson Junctions in Superconducting.pdf;/Users/wasmer/Zotero/storage/ZVNVRDSF/acs.nanolett.html}
}
@@ -14698,7 +15983,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
urldate = {2023-12-07},
abstract = {This text is designed to give the reader a helping hand in writing a scientific paper. It provides generic advice on ways that a scientific paper can be improved. The focus is on the following ethical and non-technical issues: (1) when to start writing, and in what language; (2) how to choose a good title; (3) what should be included in the various sections (abstract, introduction, experimental, results, discussion, conclusions, and supporting information (supplementary material); (4) who should be considered as a co-author, and who should be acknowledged for help; (5) which journal should be chosen; and (6) how to respond to reviewers’ comments. Purely technical issues, such as grammar, artwork, reference styles, etc., are not considered.},
langid = {english},
- keywords = {/unread,\_tablet,advice,best practices,educational,publishing,scientific journals,scientific writing,working in science,writing},
+ keywords = {/unread,advice,best practices,educational,publishing,scientific journals,scientific writing,working in science,writing},
file = {/Users/wasmer/Nextcloud/Zotero/Scholz_2022_Writing and publishing a scientific paper.pdf}
}
@@ -14765,17 +16050,36 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Schütt, Kristof T. and Hessmann, Stefaan S. P. and Gebauer, Niklas W. A. and Lederer, Jonas and Gastegger, Michael},
date = {2022-12-11},
eprint = {2212.05517},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2212.05517},
url = {http://arxiv.org/abs/2212.05517},
urldate = {2022-12-27},
abstract = {SchNetPack is a versatile neural networks toolbox that addresses both the requirements of method development and application of atomistic machine learning. Version 2.0 comes with an improved data pipeline, modules for equivariant neural networks as well as a PyTorch implementation of molecular dynamics. An optional integration with PyTorch Lightning and the Hydra configuration framework powers a flexible command-line interface. This makes SchNetPack 2.0 easily extendable with custom code and ready for complex training task such as generation of 3d molecular structures.},
- pubstate = {preprint},
- keywords = {\_tablet,Deep learning,equivariant,Hydra,library,MLP,models,PAiNN,pytorch,SchNet,SO(3),with-code},
+ pubstate = {prepublished},
+ keywords = {/unread,equivariant,Hydra,library,MLP,models,PAiNN,pytorch,pytorch lightning,SchNet,SchNetPack,SO(3),with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Schütt et al_2022_SchNetPack 2.pdf;/Users/wasmer/Zotero/storage/AHBKQSBM/2212.html}
}
+@article{schuttSchNetPackNeuralNetwork2023,
+ title = {{{SchNetPack}} 2.0: {{A}} Neural Network Toolbox for Atomistic Machine Learning},
+ shorttitle = {{{SchNetPack}} 2.0},
+ author = {Schütt, Kristof T. and Hessmann, Stefaan S. P. and Gebauer, Niklas W. A. and Lederer, Jonas and Gastegger, Michael},
+ date = {2023-04-12},
+ journaltitle = {The Journal of Chemical Physics},
+ shortjournal = {The Journal of Chemical Physics},
+ volume = {158},
+ number = {14},
+ pages = {144801},
+ issn = {0021-9606},
+ doi = {10.1063/5.0138367},
+ url = {https://doi.org/10.1063/5.0138367},
+ urldate = {2024-07-04},
+ abstract = {SchNetPack is a versatile neural network toolbox that addresses both the requirements of method development and the application of atomistic machine learning. Version 2.0 comes with an improved data pipeline, modules for equivariant neural networks, and a PyTorch implementation of molecular dynamics. An optional integration with PyTorch Lightning and the Hydra configuration framework powers a flexible command-line interface. This makes SchNetPack 2.0 easily extendable with a custom code and ready for complex training tasks, such as the generation of 3D molecular structures.},
+ keywords = {/unread,equivariant,Hydra,library,MLP,models,PAiNN,pytorch,pytorch lightning,SchNet,SchNetPack,SO(3),with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Schütt et al_2023_SchNetPack 2.pdf;/Users/wasmer/Zotero/storage/N36DDUS9/SchNetPack-2-0-A-neural-network-toolbox-for.html}
+}
+
@article{schuttUnifyingMachineLearning2019,
title = {Unifying Machine Learning and Quantum Chemistry with a Deep Neural Network for Molecular Wavefunctions},
author = {Schütt, K. T. and Gastegger, M. and Tkatchenko, A. and Müller, K.-R. and Maurer, R. J.},
@@ -14843,7 +16147,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
urldate = {2022-07-03},
abstract = {Machine learning (ML)-based approaches to battery design are relatively new but demonstrate significant promise for accelerating the timeline for new materials discovery, process optimization, and cell lifetime prediction. Battery modeling represents an interesting and unconventional application area for ML, as datasets are often small but some degree of physical understanding of the underlying processes may exist. This review article provides discussion and analysis of several important and increasingly common questions: how ML-based battery modeling works, how much data are required, how to judge model performance, and recommendations for building models in the small data regime. This article begins with an introduction to ML in general, highlighting several important concepts for small data applications. Previous ionic conductivity modeling efforts are discussed in depth as a case study to illustrate these modeling concepts. Finally, an overview of modeling efforts in major areas of battery design is provided and several areas for promising future efforts are identified, within the context of typical small data constraints.},
langid = {english},
- keywords = {\_tablet,chemistry,materials informatics,ML,small data,tutorial},
+ keywords = {chemistry,materials informatics,ML,small data,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/Sendek et al_Machine Learning Modeling for Accelerated Battery Materials Design in the Small.pdf;/Users/wasmer/Zotero/storage/55KE647F/aenm.html}
}
@@ -14852,13 +16156,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Sevilla, Jaime and Heim, Lennart and Ho, Anson and Besiroglu, Tamay and Hobbhahn, Marius and Villalobos, Pablo},
date = {2022-03-09},
eprint = {2202.05924},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2202.05924},
url = {http://arxiv.org/abs/2202.05924},
urldate = {2023-09-19},
abstract = {Compute, data, and algorithmic advances are the three fundamental factors that guide the progress of modern Machine Learning (ML). In this paper we study trends in the most readily quantified factor - compute. We show that before 2010 training compute grew in line with Moore's law, doubling roughly every 20 months. Since the advent of Deep Learning in the early 2010s, the scaling of training compute has accelerated, doubling approximately every 6 months. In late 2015, a new trend emerged as firms developed large-scale ML models with 10 to 100-fold larger requirements in training compute. Based on these observations we split the history of compute in ML into three eras: the Pre Deep Learning Era, the Deep Learning Era and the Large-Scale Era. Overall, our work highlights the fast-growing compute requirements for training advanced ML systems.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AI,ecological footprint,economics,energy consumption,energy efficiency,environmental impact,for introductions,General ML,history of AI,history of science,HPC,ML,Moore's Law,Our World in Data,policy,supercomputing},
file = {/Users/wasmer/Nextcloud/Zotero/Sevilla et al_2022_Compute Trends Across Three Eras of Machine Learning.pdf;/Users/wasmer/Zotero/storage/WA244FAW/2202.html}
}
@@ -14868,13 +16172,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Shah, Karan and Stiller, Patrick and Hoffmann, Nico and Cangi, Attila},
date = {2022-10-22},
eprint = {2210.12522},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {quant-ph},
doi = {10.48550/arXiv.2210.12522},
url = {http://arxiv.org/abs/2210.12522},
urldate = {2023-02-15},
abstract = {We demonstrate the utility of physics-informed neural networks (PINNs) as solvers for the non-relativistic, time-dependent Schr\textbackslash "odinger equation. We study the performance and generalisability of PINN solvers on the time evolution of a quantum harmonic oscillator across varying system parameters, domains, and energy states.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread},
file = {/Users/wasmer/Nextcloud/Zotero/Shah et al_2022_Physics-Informed Neural Networks as Solvers for the Time-Dependent.pdf;/Users/wasmer/Zotero/storage/NSJSIKTH/2210.html}
}
@@ -14909,7 +16213,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
urldate = {2024-03-08},
abstract = {The theorems of density functional theory (DFT) establish bijective maps between the local external potential of a many-body system and its electron density, wavefunction and, therefore, one-particle reduced density matrix. Building on this foundation, we show that machine learning models based on the one-electron reduced density matrix can be used to generate surrogate electronic structure methods. We generate surrogates of local and hybrid DFT, Hartree-Fock and full configuration interaction theories for systems ranging from small molecules such as water to more complex compounds like benzene and propanol. The surrogate models use the one-electron reduced density matrix as the central quantity to be learned. From the predicted density matrices, we show that either standard quantum chemistry or a second machine-learning model can be used to compute molecular observables, energies, and atomic forces. The surrogate models can generate essentially anything that a standard electronic structure method can, ranging from band gaps and Kohn-Sham orbitals to energy-conserving ab-initio molecular dynamics simulations and infrared spectra, which account for anharmonicity and thermal effects, without the need to employ computationally expensive algorithms such as self-consistent field theory. The algorithms are packaged in an efficient and easy to use Python code, QMLearn, accessible on popular platforms.},
langid = {english},
- keywords = {\_tablet,AML,ase,B3LYP,configuration interaction,density matrix,DFT,DFT speedup,DFT speedup with ML,GTO basis,HFT,KRR,LDA,library,MD,ML,ML-DFT,ML-ESM,ML-WFT,molecules,multi-step model,prediction from potential,prediction of density matrix,prediction of electron density,prediction of energy,prediction of forces,PySCF,RDMFT,SCF,surrogate model,WFT,with-code,with-data},
+ keywords = {AML,ase,B3LYP,configuration interaction,density matrix,DFT,DFT speedup,DFT speedup with ML,GTO basis,HFT,KRR,LDA,library,MD,ML,ML-DFT,ML-ESM,ML-WFT,molecules,multi-step model,prediction from potential,prediction of density matrix,prediction of electron density,prediction of energy,prediction of forces,PySCF,RDMFT,SCF,surrogate model,WFT,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Shao et al_2023_Machine learning electronic structure methods based on the one-electron reduced.pdf}
}
@@ -14927,7 +16231,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
urldate = {2023-04-04},
abstract = {In this paper I propose a new model for representing the formation energies of multicomponent crystalline alloys as a function of atom types. In the cases when displacements of atoms from their equilibrium positions are not large, the proposed method has a similar accuracy as the state-of-the-art cluster expansion method, and a better accuracy when the fitting dataset size is small. The proposed model has only two tunable parameters—one for the interaction range and one for the interaction complexity.},
langid = {english},
- keywords = {\_tablet,alloys,AML,cluster expansion,high-entropy alloys,LRP,ML,MLP,n-ary alloys,original publication,prediction of energy,transition metals},
+ keywords = {alloys,AML,cluster expansion,high-entropy alloys,LRP,ML,MLP,n-ary alloys,original publication,prediction of energy,transition metals},
file = {/Users/wasmer/Nextcloud/Zotero/Shapeev_2017_Accurate representation of formation energies of crystalline alloys with many.pdf;/Users/wasmer/Zotero/storage/EQYE3F3F/S0927025617303610.html}
}
@@ -14936,14 +16240,14 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Shen, Jimmy-Xuan and Munro, Jason M. and Horton, Matthew K. and Huck, Patrick and Dwaraknath, Shyam and Persson, Kristin A.},
date = {2021-07-07},
eprint = {2107.03540},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2107.03540},
url = {http://arxiv.org/abs/2107.03540},
urldate = {2022-12-31},
abstract = {In addition to being the core quantity in density functional theory, the charge density can be used in many tertiary analyses in materials sciences from bonding to assigning charge to specific atoms. The charge density is data-rich since it contains information about all the electrons in the system. With increasing utilization of machine-learning tools in materials sciences, a data-rich object like the charge density can be utilized in a wide range of applications. The database presented here provides a modern and user-friendly interface for a large and continuously updated collection of charge densities as part of the Materials Project. In addition to the charge density data, we provide the theory and code for changing the representation of the charge density which should enable more advanced machine-learning studies for the broader community.},
- pubstate = {preprint},
- keywords = {\_tablet,charge density,data repositories,Database,dimensionality reduction of target,electronic structure,grid-based descriptors,library,materials,materials database,materials project,ML,ML-DFT,prediction from density,prediction of electron density,pseudopotential,representation of density,VASP,with-code},
+ pubstate = {prepublished},
+ keywords = {charge density,data repositories,Database,dimensionality reduction of target,electronic structure,grid-based descriptors,library,materials,materials database,materials project,ML,ML-DFT,prediction from density,prediction of electron density,pseudopotential,representation of density,VASP,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Shen et al_2021_A representation-independent electronic charge density database for crystalline.pdf;/Users/wasmer/Zotero/storage/9A3MUVVK/2107.html}
}
@@ -14952,17 +16256,35 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Shen, Paul and Herbst, Michael and Viswanathan, Venkat},
date = {2022-08-04},
eprint = {2108.09541},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2108.09541},
url = {http://arxiv.org/abs/2108.09541},
urldate = {2023-05-26},
abstract = {We develop theory and software for rotation equivariant operators on scalar and vector fields, with diverse applications in simulation, optimization and machine learning. Rotation equivariance (covariance) means all fields in the system rotate together, implying spatially invariant dynamics that preserve symmetry. Extending the convolution theorems of linear time invariant systems, we theorize that linear equivariant operators are characterized by tensor field convolutions using an appropriate product between the input field and a radially symmetric kernel field. Most Green's functions and differential operators are in fact equivariant operators, which can also fit unknown symmetry preserving dynamics by parameterizing the radial function. We implement the Julia package EquivariantOperators.jl for fully differentiable finite difference equivariant operators on scalar, vector and higher order tensor fields in 2d/3d. It can run forwards for simulation or image processing, or be back propagated for computer vision, inverse problems and optimal control. Code at https://aced-differentiate.github.io/EquivariantOperators.jl/},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,CNN,covariant,differential operator,electric charge,Electric field,electric potential,equivariant,equivariant operator,finite differences,General ML,Green’s Function Method,grid-based descriptors,image classification,Julia,library,ML,ML-DFA,ML-DFT,ML-ESM,neural nets,nonuniform grid,operator,PDE,PINN,prediction from density,prediction of,prediction of Exc,scalar field,tensor field,tensorial target,vector field,vectorial learning target,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Shen et al_2022_Rotation Equivariant Operators for Machine Learning on Scalar and Vector Fields.pdf;/Users/wasmer/Zotero/storage/NYTWYPMA/2108.html}
}
+@article{shenSimulatingChargedDefects2024,
+ title = {Simulating Charged Defects at Database Scale},
+ author = {Shen, Jimmy-Xuan and Voss, Lars F. and Varley, Joel B.},
+ date = {2024-04-10},
+ journaltitle = {Journal of Applied Physics},
+ shortjournal = {Journal of Applied Physics},
+ volume = {135},
+ number = {14},
+ pages = {145102},
+ issn = {0021-8979},
+ doi = {10.1063/5.0203124},
+ url = {https://doi.org/10.1063/5.0203124},
+ urldate = {2024-05-24},
+ abstract = {Point defects have a strong influence on the physical properties of materials, often dominating the electronic and optical behavior in semiconductors and insulators. The simulation and analysis of point defects is, therefore, crucial for understanding the growth and operation of materials, especially for optoelectronics applications. In this work, we present a general-purpose Python framework for the analysis of point defects in crystalline materials as well as a generalized workflow for their treatment with high-throughput simulations. The distinguishing feature of our approach is an emphasis on a unique, unit cell, structure-only, definition of point defects which decouples the defect definition, and the specific supercell representation used to simulate the defect. This allows the results of first-principles calculations to be aggregated into a database without extensive provenance information and is a crucial step in building a persistent database of point defects that can grow over time, a key component toward realizing the idea of a “defect genome†that can yield more complex relationships governing the behavior of defects in materials. We demonstrate several examples of the approach for three technologically relevant materials and highlight current pitfalls that must be considered when employing these methodologies as well as their potential solutions.},
+ keywords = {atomate,database generation,defects,DFT,HTC,library,materials,physics,point defects,pymatgen,Python,supercell,with-code,workflows},
+ file = {/Users/wasmer/Nextcloud/Zotero/Shen et al_2024_Simulating charged defects at database scale.pdf;/Users/wasmer/Zotero/storage/LV65KWLW/Simulating-charged-defects-at-database-scale.html}
+}
+
@article{shibaClassicalSpinsSuperconductors1968,
title = {Classical {{Spins}} in {{Superconductors}}},
author = {Shiba, Hiroyuki},
@@ -14986,13 +16308,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Shirobokov, Sergey and Belavin, Vladislav and Kagan, Michael and Ustyuzhanin, Andrey and Baydin, Atılım Güneş},
date = {2020-06-15},
eprint = {2002.04632},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {hep-ex, physics:physics, stat},
doi = {10.48550/arXiv.2002.04632},
url = {http://arxiv.org/abs/2002.04632},
urldate = {2023-11-18},
abstract = {We propose a novel method for gradient-based optimization of black-box simulators using differentiable local surrogate models. In fields such as physics and engineering, many processes are modeled with non-differentiable simulators with intractable likelihoods. Optimization of these forward models is particularly challenging, especially when the simulator is stochastic. To address such cases, we introduce the use of deep generative models to iteratively approximate the simulator in local neighborhoods of the parameter space. We demonstrate that these local surrogates can be used to approximate the gradient of the simulator, and thus enable gradient-based optimization of simulator parameters. In cases where the dependence of the simulator on the parameter space is constrained to a low dimensional submanifold, we observe that our method attains minima faster than baseline methods, including Bayesian optimization, numerical optimization, and approaches using score function gradient estimators.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {active learning,AI4Science,alternative approaches,autodiff,Bayesian optimization,differentiable programming,Gaussian process,high-energy physics,hybrid AI/simulation,library,ML,optimization,particle physics,surrogate model,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Shirobokov et al_2020_Black-Box Optimization with Local Generative Surrogates.pdf;/Users/wasmer/Zotero/storage/E8SUJGVP/2002.html}
}
@@ -15010,7 +16332,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
url = {https://doi.org/10.1021/acs.jctc.2c00555},
urldate = {2022-09-29},
abstract = {Leveraging ab initio data at scale has enabled the development of machine learning models capable of extremely accurate and fast molecular property prediction. A central paradigm of many previous studies focuses on generating predictions for only a fixed set of properties. Recent lines of research instead aim to explicitly learn the electronic structure via molecular wavefunctions, from which other quantum chemical properties can be directly derived. While previous methods generate predictions as a function of only the atomic configuration, in this work we present an alternate approach that directly purposes basis-dependent information to predict molecular electronic structure. Our model, Orbital Mixer, is composed entirely of multi-layer perceptrons (MLPs) using MLP-Mixer layers within a simple, intuitive, and scalable architecture that achieves competitive Hamiltonian and molecular orbital energy and coefficient prediction accuracies compared to the state-of-the-art.},
- keywords = {\_tablet,ML,ML-ESM,MLP,molecules,Orbital Mixer,original publication,PhiSNet,prediction of wavefunction,SchNOrb,with-code},
+ keywords = {ML,ML-ESM,MLP,molecules,Orbital Mixer,original publication,PhiSNet,prediction of wavefunction,SchNOrb,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Shmilovich et al_2022_Orbital Mixer.pdf}
}
@@ -15020,29 +16342,46 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Shoghi, Nima and Kolluru, Adeesh and Kitchin, John R. and Ulissi, Zachary W. and Zitnick, C. Lawrence and Wood, Brandon M.},
date = {2024-05-06},
eprint = {2310.16802},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2310.16802},
url = {http://arxiv.org/abs/2310.16802},
urldate = {2024-05-07},
abstract = {Foundation models have been transformational in machine learning fields such as natural language processing and computer vision. Similar success in atomic property prediction has been limited due to the challenges of training effective models across multiple chemical domains. To address this, we introduce Joint Multi-domain Pre-training (JMP), a supervised pre-training strategy that simultaneously trains on multiple datasets from different chemical domains, treating each dataset as a unique pre-training task within a multi-task framework. Our combined training dataset consists of \$\textbackslash sim\$120M systems from OC20, OC22, ANI-1x, and Transition-1x. We evaluate performance and generalization by fine-tuning over a diverse set of downstream tasks and datasets including: QM9, rMD17, MatBench, QMOF, SPICE, and MD22. JMP demonstrates an average improvement of 59\% over training from scratch, and matches or sets state-of-the-art on 34 out of 40 tasks. Our work highlights the potential of pre-training strategies that utilize diverse data to advance property prediction across chemical domains, especially for low-data tasks. Please visit https://nima.sh/jmp for further information.},
- pubstate = {preprint},
- keywords = {Allegro,AML,ANI1-x,benchmarking,fine-tuning,foundation models,JMP,MACE,MatBench,MD17,Meta Research,ML,MLP,MODNet,multi-domain,multi-task learning,OC20,Open Catalyst,pretrained models,QM9,sGDML,SPICE dataset,universal potential},
+ pubstate = {prepublished},
+ keywords = {Allegro,AML,ANI1-x,benchmarking,fine-tuning,foundation models,JMP,MACE,MatBench,MD17,Meta Research,ML,MLP,MODNet,multi-domain,multi-task learning,OC20,Open Catalyst,pretrained models,QM9,sGDML,SPICE dataset,universal potential,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Shoghi et al_2024_From Molecules to Materials.pdf;/Users/wasmer/Zotero/storage/XXHLMYBL/2310.html}
}
+@online{shuForecastingFutureFuture2024,
+ title = {Forecasting the {{Future}} with {{Future Technologies}}: {{Advancements}} in {{Large Meteorological Models}}},
+ shorttitle = {Forecasting the {{Future}} with {{Future Technologies}}},
+ author = {Shu, Hailong and Wang, Yue and Song, Weiwei and Guo, Huichuang and Song, Zhen},
+ date = {2024-04-09},
+ eprint = {2404.06668},
+ eprinttype = {arXiv},
+ eprintclass = {physics},
+ doi = {10.48550/arXiv.2404.06668},
+ url = {http://arxiv.org/abs/2404.06668},
+ urldate = {2024-08-02},
+ abstract = {The field of meteorological forecasting has undergone a significant transformation with the integration of large models, especially those employing deep learning techniques. This paper reviews the advancements and applications of these models in weather prediction, emphasizing their role in transforming traditional forecasting methods. Models like FourCastNet, Pangu-Weather, GraphCast, ClimaX, and FengWu have made notable contributions by providing accurate, high-resolution forecasts, surpassing the capabilities of traditional Numerical Weather Prediction (NWP) models. These models utilize advanced neural network architectures, such as Convolutional Neural Networks (CNNs), Graph Neural Networks (GNNs), and Transformers, to process diverse meteorological data, enhancing predictive accuracy across various time scales and spatial resolutions. The paper addresses challenges in this domain, including data acquisition and computational demands, and explores future opportunities for model optimization and hardware advancements. It underscores the integration of artificial intelligence with conventional meteorological techniques, promising improved weather prediction accuracy and a significant contribution to addressing climate-related challenges. This synergy positions large models as pivotal in the evolving landscape of meteorological forecasting.},
+ pubstate = {prepublished},
+ keywords = {/unread,AI4Science,foundation models,large models,review,weather forecasting},
+ file = {/Users/wasmer/Nextcloud/Zotero/Shu et al_2024_Forecasting the Future with Future Technologies.pdf;/Users/wasmer/Zotero/storage/ELHX63Z2/2404.html}
+}
+
@online{simeonInclusionChargeSpin2024,
title = {On the {{Inclusion}} of {{Charge}} and {{Spin States}} in {{Cartesian Tensor Neural Network Potentials}}},
author = {Simeon, Guillem and Mirarchi, Antonio and Pelaez, Raul P. and Galvelis, Raimondas and De Fabritiis, Gianni},
date = {2024-03-22},
eprint = {2403.15073},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2403.15073},
url = {http://arxiv.org/abs/2403.15073},
urldate = {2024-03-31},
abstract = {In this letter, we present an extension to TensorNet, a state-of-the-art equivariant Cartesian tensor neural network potential, allowing it to handle charged molecules and spin states without architectural changes or increased costs. By incorporating these attributes, we address input degeneracy issues, enhancing the model's predictive accuracy across diverse chemical systems. This advancement significantly broadens TensorNet's applicability, maintaining its efficiency and accuracy.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AML,equivariant,ML,MLP,molecules,prediction of charge,prediction of energy,prediction of spin state,QMSpin,SPICE dataset,spin,spin-dependent,TensorNet,TorchMDNet},
file = {/Users/wasmer/Nextcloud/Zotero/Simeon et al_2024_On the Inclusion of Charge and Spin States in Cartesian Tensor Neural Network.pdf;/Users/wasmer/Zotero/storage/U8FJBM8P/2403.html}
}
@@ -15098,7 +16437,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
langid = {english},
pagetotal = {134},
keywords = {DFT,DFT theory,educational,FLAPW,LAPW,learn DFT,magnetism,numerical,physics,textbook},
- file = {/Users/wasmer/Nextcloud/Zotero/2006_Planewaves, Pseudopotentials and the LAPW Method.pdf}
+ file = {/Users/wasmer/Nextcloud/Zotero/Singh_Nordström_2006_Planewaves, Pseudopotentials and the LAPW Method.pdf}
}
@article{singhRareearthBasedHalfHeusler2020,
@@ -15162,13 +16501,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Sinitskiy, Anton V. and Pande, Vijay S.},
date = {2018-09-07},
eprint = {1809.02723},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.1809.02723},
url = {http://arxiv.org/abs/1809.02723},
urldate = {2023-10-01},
abstract = {Density functional theory (DFT) is one of the main methods in Quantum Chemistry that offers an attractive trade off between the cost and accuracy of quantum chemical computations. The electron density plays a key role in DFT. In this work, we explore whether machine learning - more specifically, deep neural networks (DNNs) - can be trained to predict electron densities faster than DFT. First, we choose a practically efficient combination of a DFT functional and a basis set (PBE0/pcS-3) and use it to generate a database of DFT solutions for more than 133,000 organic molecules from a previously published database QM9. Next, we train a DNN to predict electron densities and energies of such molecules. The only input to the DNN is an approximate electron density computed with a cheap quantum chemical method in a small basis set (HF/cc-VDZ). We demonstrate that the DNN successfully learns differences in the electron densities arising both from electron correlation and small basis set artifacts in the HF computations. All qualitative features in density differences, including local minima on lone pairs, local maxima on nuclei, toroidal shapes around C-H and C-C bonds, complex shapes around aromatic and cyclopropane rings and CN group, etc. are captured by the DNN. Accuracy of energy predictions by the DNN is \textasciitilde{} 1 kcal/mol, on par with other models reported in the literature, while those models do not predict the electron density. Computations with the DNN, including HF computations, take much less time that DFT computations (by a factor of \textasciitilde 20-30 for most QM9 molecules in the current version, and it is clear how it could be further improved).},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,benchmarking,CCSD(T),CNN,delta learning,ML,ML-DFT,ML-ESM,MLP,MLP comparison,molecules,organic chemistry,PBE,prediction of electron density,prediction of energy,QM9,U-net,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Sinitskiy_Pande_2018_Deep Neural Network Computes Electron Densities and Energies of a Large Set of.pdf;/Users/wasmer/Zotero/storage/C263WSGH/1809.html}
}
@@ -15244,7 +16583,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
url = {https://link.aps.org/doi/10.1103/PhysRevLett.108.253002},
urldate = {2021-10-15},
abstract = {Machine learning is used to approximate density functionals. For the model problem of the kinetic energy of noninteracting fermions in 1D, mean absolute errors below 1 kcal/mol on test densities similar to the training set are reached with fewer than 100 training densities. A predictor identifies if a test density is within the interpolation region. Via principal component analysis, a projected functional derivative finds highly accurate self-consistent densities. The challenges for application of our method to real electronic structure problems are discussed.},
- keywords = {\_tablet,2-step model,DFT,dimensionality reduction,KRR,ML,ML-DFA,ML-DFT,ML-ESM,ML-OF,models,orbital-free DFT,original publication,PCA,prediction from density,prediction from density functional,prediction of electron density,prediction of kinetic energy},
+ keywords = {2-step model,DFT,dimensionality reduction,KRR,ML,ML-DFA,ML-DFT,ML-ESM,ML-OF,models,orbital-free DFT,original publication,PCA,prediction from density,prediction from density functional,prediction of electron density,prediction of kinetic energy},
file = {/Users/wasmer/Nextcloud/Zotero/Snyder et al_2012_Finding Density Functionals with Machine Learning.pdf;/Users/wasmer/Zotero/storage/6NMNCTQB/Snyder et al. - 2012 - Finding Density Functionals with Machine Learning.tex;/Users/wasmer/Zotero/storage/TBZPF93I/Snyder et al_2012_Finding Density Functionals with Machine Learning.pdf;/Users/wasmer/Zotero/storage/RRS5SC4P/PhysRevLett.108.html}
}
@@ -15284,8 +16623,8 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
abstract = {Data-driven methods, in particular machine learning, can help to speed up the discovery of new materials by finding hidden patterns in existing data and using them to identify promising candidate materials. In the case of superconductors, the use of data science tools is to date slowed down by a lack of accessible data. In this work, we present a new and publicly available superconductivity dataset (‘3DSC’), featuring the critical temperature TC of superconducting materials additionally to tested non-superconductors. In contrast to existing databases such as the SuperCon database which contains information on the chemical composition, the 3DSC is augmented by approximate three-dimensional crystal structures. We perform a statistical analysis and machine learning experiments to show that access to this structural information improves the prediction of the critical temperature TC of materials. Furthermore, we provide ideas and directions for further research to improve the 3DSC. We are confident that this database will be useful in applying state-of-the-art machine learning methods to eventually find new superconductors.},
issue = {1},
langid = {english},
- keywords = {/unread,AML,Database,descriptors,disordered,disordered SOAP,ensemble learning,magpie,ML,SOAP,superconductor,with-code,with-data},
- file = {/Users/wasmer/Nextcloud/Zotero/Sommer et al_2023_3DSC - a dataset of superconductors including crystal structures.pdf}
+ keywords = {AML,Database,defects,descriptors,disordered,disordered SOAP,ensemble learning,magpie,ML,SOAP,superconductor,with-code,with-data},
+ file = {/Users/wasmer/Nextcloud/Zotero/Sommer et al_2023_3DSC - a dataset of superconductors including crystal structures.pdf;/Users/wasmer/Nextcloud/Zotero/Sommer et al_2023_3DSC - a dataset of superconductors including crystal structures3.pdf}
}
@online{sommer3DSCNewDataset2022,
@@ -15293,14 +16632,14 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Sommer, Timo and Willa, Roland and Schmalian, Jörg and Friederich, Pascal},
date = {2022-12-14},
eprint = {2212.06071},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2212.06071},
url = {http://arxiv.org/abs/2212.06071},
urldate = {2023-02-15},
abstract = {Data-driven methods, in particular machine learning, can help to speed up the discovery of new materials by finding hidden patterns in existing data and using them to identify promising candidate materials. In the case of superconductors, which are a highly interesting but also a complex class of materials with many relevant applications, the use of data science tools is to date slowed down by a lack of accessible data. In this work, we present a new and publicly available superconductivity dataset ('3DSC'), featuring the critical temperature \$T\_\textbackslash mathrm\{c\}\$ of superconducting materials additionally to tested non-superconductors. In contrast to existing databases such as the SuperCon database which contains information on the chemical composition, the 3DSC is augmented by the approximate three-dimensional crystal structure of each material. We perform a statistical analysis and machine learning experiments to show that access to this structural information improves the prediction of the critical temperature \$T\_\textbackslash mathrm\{c\}\$ of materials. Furthermore, we see the 3DSC not as a finished dataset, but we provide ideas and directions for further research to improve the 3DSC in multiple ways. We are confident that this database will be useful in applying state-of-the-art machine learning methods to eventually find new superconductors.},
- pubstate = {preprint},
- keywords = {AML,Database,descriptors,disordered,disordered SOAP,ensemble learning,magpie,ML,SOAP,superconductor,with-code,with-data},
+ pubstate = {prepublished},
+ keywords = {AML,Database,defects,descriptors,disordered,disordered SOAP,ensemble learning,magpie,ML,SOAP,superconductor,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Sommer et al_2022_3DSC - A New Dataset of Superconductors Including Crystal Structures.pdf;/Users/wasmer/Zotero/storage/JMMVYJCI/2212.html}
}
@@ -15310,13 +16649,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Song, Shuaiwen Leon and Kruft, Bonnie and Zhang, Minjia and Li, Conglong and Chen, Shiyang and Zhang, Chengming and Tanaka, Masahiro and Wu, Xiaoxia and Rasley, Jeff and Awan, Ammar Ahmad and Holmes, Connor and Cai, Martin and Ghanem, Adam and Zhou, Zhongzhu and He, Yuxiong and Luferenko, Pete and Kumar, Divya and Weyn, Jonathan and Zhang, Ruixiong and Klocek, Sylwester and Vragov, Volodymyr and AlQuraishi, Mohammed and Ahdritz, Gustaf and Floristean, Christina and Negri, Cristina and Kotamarthi, Rao and Vishwanath, Venkatram and Ramanathan, Arvind and Foreman, Sam and Hippe, Kyle and Arcomano, Troy and Maulik, Romit and Zvyagin, Maxim and Brace, Alexander and Zhang, Bin and Bohorquez, Cindy Orozco and Clyde, Austin and Kale, Bharat and Perez-Rivera, Danilo and Ma, Heng and Mann, Carla M. and Irvin, Michael and Pauloski, J. Gregory and Ward, Logan and Hayot, Valerie and Emani, Murali and Xie, Zhen and Lin, Diangen and Shukla, Maulik and Foster, Ian and Davis, James J. and Papka, Michael E. and Brettin, Thomas and Balaprakash, Prasanna and Tourassi, Gina and Gounley, John and Hanson, Heidi and Potok, Thomas E. and Pasini, Massimiliano Lupo and Evans, Kate and Lu, Dan and Lunga, Dalton and Yin, Junqi and Dash, Sajal and Wang, Feiyi and Shankar, Mallikarjun and Lyngaas, Isaac and Wang, Xiao and Cong, Guojing and Zhang, Pei and Fan, Ming and Liu, Siyan and Hoisie, Adolfy and Yoo, Shinjae and Ren, Yihui and Tang, William and Felker, Kyle and Svyatkovskiy, Alexey and Liu, Hang and Aji, Ashwin and Dalton, Angela and Schulte, Michael and Schulz, Karl and Deng, Yuntian and Nie, Weili and Romero, Josh and Dallago, Christian and Vahdat, Arash and Xiao, Chaowei and Gibbs, Thomas and Anandkumar, Anima and Stevens, Rick},
date = {2023-10-11},
eprint = {2310.04610},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2310.04610},
url = {http://arxiv.org/abs/2310.04610},
urldate = {2023-11-05},
abstract = {In the upcoming decade, deep learning may revolutionize the natural sciences, enhancing our capacity to model and predict natural occurrences. This could herald a new era of scientific exploration, bringing significant advancements across sectors from drug development to renewable energy. To answer this call, we present DeepSpeed4Science initiative (deepspeed4science.ai) which aims to build unique capabilities through AI system technology innovations to help domain experts to unlock today's biggest science mysteries. By leveraging DeepSpeed's current technology pillars (training, inference and compression) as base technology enablers, DeepSpeed4Science will create a new set of AI system technologies tailored for accelerating scientific discoveries by addressing their unique complexity beyond the common technical approaches used for accelerating generic large language models (LLMs). In this paper, we showcase the early progress we made with DeepSpeed4Science in addressing two of the critical system challenges in structural biology research.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AI4Science,AML,Argonne National Laboratory,biomolecules,drug discovery,foundation models,GPU,HPC,library,LLM,materials discovery,Microsoft Research,ML,NVIDIA,performance optimization,Quantum chemistry,surrogate model,transformer,white paper,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Song et al_2023_DeepSpeed4Science Initiative.pdf;/Users/wasmer/Zotero/storage/9XE4R4E2/2310.html}
}
@@ -15347,13 +16686,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Song, Xingyou and Li, Oscar and Lee, Chansoo and Yang, Bangding and Peng, Daiyi and Perel, Sagi and Chen, Yutian},
date = {2024-03-04},
eprint = {2402.14547},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2402.14547},
url = {http://arxiv.org/abs/2402.14547},
urldate = {2024-03-31},
abstract = {Over the broad landscape of experimental design, regression has been a powerful tool to accurately predict the outcome metrics of a system or model given a set of parameters, but has been traditionally restricted to methods which are only applicable to a specific task. In this paper, we propose OmniPred, a framework for training language models as universal end-to-end regressors over \$(x,y)\$ evaluation data from diverse real world experiments. Using data sourced from Google Vizier, one of the largest blackbox optimization databases in the world, our extensive experiments demonstrate that through only textual representations of mathematical parameters and values, language models are capable of very precise numerical regression, and if given the opportunity to train over multiple tasks, can significantly outperform traditional regression models.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,alternative approaches,alternative for equivariance,alternative to GNN,DeepMind,GNN,Google,language models,LLM,model comparison,regression,transformer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Song et al_2024_OmniPred.pdf;/Users/wasmer/Zotero/storage/5KU8ET6U/2402.html}
}
@@ -15411,13 +16750,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Speckhard, Daniel T. and Carbogno, Christian and Ghiringhelli, Luca and Lubeck, Sven and Scheffler, Matthias and Draxl, Claudia},
date = {2023-06-01},
eprint = {2303.14760},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics, stat},
doi = {10.48550/arXiv.2303.14760},
url = {http://arxiv.org/abs/2303.14760},
urldate = {2023-10-11},
abstract = {The numerical precision of density-functional-theory (DFT) calculations depends on a variety of computational parameters, one of the most critical being the basis-set size. The ultimate precision is reached with an infinitely large basis set, i.e., in the limit of a complete basis set (CBS). Our aim in this work is to find a machine-learning model that extrapolates finite basis-size calculations to the CBS limit. We start with a data set of 63 binary solids investigated with two all-electron DFT codes, exciting and FHI-aims, which employ very different types of basis sets. A quantile-random-forest model is used to estimate the total-energy correction with respect to a fully converged calculation as a function of the basis-set size. The random-forest model achieves a symmetric mean absolute percentage error of lower than 25\% for both codes and outperforms previous approaches in the literature. Our approach also provides prediction intervals, which quantify the uncertainty of the models' predictions.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,DFT,DFT codes comparison,DFT numerics,error estimate,exciting DFT code,FHI-aims,numerical analysis,numerical errors,rec-by-bluegel},
file = {/Users/wasmer/Nextcloud/Zotero/Speckhard et al_2023_Extrapolation to complete basis-set limit in density-functional theory by.pdf;/Users/wasmer/Zotero/storage/6FSJEP2H/2303.html}
}
@@ -15427,13 +16766,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Spencer, James S. and Pfau, David and Botev, Aleksandar and Foulkes, W. M. C.},
date = {2020-11-13},
eprint = {2011.07125},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2011.07125},
url = {http://arxiv.org/abs/2011.07125},
urldate = {2022-06-25},
abstract = {The Fermionic Neural Network (FermiNet) is a recently-developed neural network architecture that can be used as a wavefunction Ansatz for many-electron systems, and has already demonstrated high accuracy on small systems. Here we present several improvements to the FermiNet that allow us to set new records for speed and accuracy on challenging systems. We find that increasing the size of the network is sufficient to reach chemical accuracy on atoms as large as argon. Through a combination of implementing FermiNet in JAX and simplifying several parts of the network, we are able to reduce the number of GPU hours needed to train the FermiNet on large systems by an order of magnitude. This enables us to run the FermiNet on the challenging transition of bicyclobutane to butadiene and compare against the PauliNet on the automerization of cyclobutadiene, and we achieve results near the state of the art for both.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {autodiff,DeepMind,FermiNet,JAX,library,MC,ML,ML-ESM,ML-QMBP,NN,PauliNet,prediction of wavefunction,QMC,VMC,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Spencer et al_2020_Better, Faster Fermionic Neural Networks.pdf;/Users/wasmer/Zotero/storage/SCSQGZ4K/2011.html}
}
@@ -15495,6 +16834,81 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
file = {/Users/wasmer/Zotero/storage/P8YS37PB/S0022103123000604.html}
}
+@article{starrettHightemperatureElectronicStructure2018,
+ title = {High-Temperature Electronic Structure with the {{Korringa-Kohn-Rostoker Green}}'s Function Method},
+ author = {Starrett, C. E.},
+ date = {2018-05-08},
+ journaltitle = {Physical Review E},
+ shortjournal = {Phys. Rev. E},
+ volume = {97},
+ number = {5},
+ pages = {053205},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevE.97.053205},
+ url = {https://link.aps.org/doi/10.1103/PhysRevE.97.053205},
+ urldate = {2024-07-05},
+ abstract = {Modeling high-temperature (tens or hundreds of eV), dense plasmas is challenging due to the multitude of non-negligible physical effects including significant partial ionization and multisite effects. These effects cause the breakdown or intractability of common methods and approximations used at low temperatures, such as pseudopotentials or plane-wave basis sets. Here we explore the Korringa-Kohn-Rostoker Green's function method at these high-temperature conditions. The method is all electron, does not rely on pseudopotentials, and uses a spherical harmonic basis set, and so avoids the aforementioned limitations. It is found to be accurate for solid density aluminum and iron plasmas when compared to a plane-wave method at low temperature, while being able to access high temperatures.},
+ keywords = {/unread},
+ file = {/Users/wasmer/Nextcloud/Zotero/Starrett_2018_High-temperature electronic structure with the Korringa-Kohn-Rostoker Green's.pdf;/Users/wasmer/Zotero/storage/MHGRYKXI/PhysRevE.97.html}
+}
+
+@article{stefanouCalculationShapetruncationFunctions1991,
+ title = {Calculation of Shape-Truncation Functions for {{Voronoi}} Polyhedra},
+ author = {Stefanou, N. and Zeller, R.},
+ date = {1991-09},
+ journaltitle = {Journal of Physics: Condensed Matter},
+ shortjournal = {J. Phys.: Condens. Matter},
+ volume = {3},
+ number = {39},
+ pages = {7599},
+ issn = {0953-8984},
+ doi = {10.1088/0953-8984/3/39/006},
+ url = {https://dx.doi.org/10.1088/0953-8984/3/39/006},
+ urldate = {2024-07-05},
+ abstract = {The authors develop a new efficient method for calculating the shape-truncation functions of arbitrary Voronoi polyhedra by combining analytical and numerical techniques. Applications are presented for cells of cubic symmetry as well as for hexagonal close-packed (HCP) atomic polyhedra with different values of the c/a ratio. They also discuss an efficient way for performing three-dimensional integrations in electronic-structure calculations (e.g. solve Poisson's equation) using shape functions.},
+ langid = {english},
+ keywords = {condensed matter,DFT,Electronic structure,JuKKR,KKR,PGI,PGI-1/IAS-1,physics,voronoi tessellation},
+ file = {/Users/wasmer/Nextcloud/Zotero/Stefanou_Zeller_1991_Calculation of shape-truncation functions for Voronoi polyhedra.pdf}
+}
+
+@article{stefanouEfficientNumericalMethod1990,
+ title = {An Efficient Numerical Method to Calculate Shape Truncation Functions for {{Wigner-Seitz}} Atomic Polyhedra},
+ author = {Stefanou, N. and Akai, H. and Zeller, R.},
+ date = {1990-09-01},
+ journaltitle = {Computer Physics Communications},
+ shortjournal = {Computer Physics Communications},
+ volume = {60},
+ number = {2},
+ pages = {231--238},
+ issn = {0010-4655},
+ doi = {10.1016/0010-4655(90)90009-P},
+ url = {https://www.sciencedirect.com/science/article/pii/001046559090009P},
+ urldate = {2024-07-05},
+ abstract = {We present an efficient numerical method to calculate shape truncation functions for Wigner-Seitz atomic polyhedra in crystalline solids. We apply our method to the case of simple cubic (SC), face centered cubic (FCC) and body centered cubic (BCC) lattices. A straightforward comparison of our results and of those obtained by other numerical techniques with the exact ones, analytically known in some simple cases, shows the efficiency and high accuracy of our method.},
+ keywords = {condensed matter,DFT,Electronic structure,JuKKR,KKR,PGI,PGI-1/IAS-1,physics,voronoi tessellation},
+ file = {/Users/wasmer/Nextcloud/Zotero/Stefanou et al_1990_An efficient numerical method to calculate shape truncation functions for.pdf;/Users/wasmer/Zotero/storage/D26W7QAE/001046559090009P.html}
+}
+
+@article{steffenTrajectoryAnthropoceneGreat2015,
+ title = {The Trajectory of the {{Anthropocene}}: {{The Great Acceleration}}},
+ shorttitle = {The Trajectory of the {{Anthropocene}}},
+ author = {Steffen, Will and Broadgate, Wendy and Deutsch, Lisa and Gaffney, Owen and Ludwig, Cornelia},
+ date = {2015-04-01},
+ journaltitle = {The Anthropocene Review},
+ volume = {2},
+ number = {1},
+ pages = {81--98},
+ publisher = {SAGE Publications},
+ issn = {2053-0196},
+ doi = {10.1177/2053019614564785},
+ url = {https://doi.org/10.1177/2053019614564785},
+ urldate = {2024-08-01},
+ abstract = {The ‘Great Acceleration’ graphs, originally published in 2004 to show socio-economic and Earth System trends from 1750 to 2000, have now been updated to 2010. In the graphs of socio-economic trends, where the data permit, the activity of the wealthy (OECD) countries, those countries with emerging economies, and the rest of the world have now been differentiated. The dominant feature of the socio-economic trends is that the economic activity of the human enterprise continues to grow at a rapid rate. However, the differentiated graphs clearly show that strong equity issues are masked by considering global aggregates only. Most of the population growth since 1950 has been in the non-OECD world but the world’s economy (GDP), and hence consumption, is still strongly dominated by the OECD world. The Earth System indicators, in general, continued their long-term, post-industrial rise, although a few, such as atmospheric methane concentration and stratospheric ozone loss, showed a slowing or apparent stabilisation over the past decade. The post-1950 acceleration in the Earth System indicators remains clear. Only beyond the mid-20th century is there clear evidence for fundamental shifts in the state and functioning of the Earth System that are beyond the range of variability of the Holocene and driven by human activities. Thus, of all the candidates for a start date for the Anthropocene, the beginning of the Great Acceleration is by far the most convincing from an Earth System science perspective.},
+ langid = {english},
+ keywords = {anthropocene,climate change,energy challenge,energy consumption,for introductions,great acceleration},
+ file = {/Users/wasmer/Nextcloud/Zotero/Steffen et al_2015_The trajectory of the Anthropocene.pdf}
+}
+
@unpublished{steinbachReproducibilityDataScience2022,
type = {presentation},
title = {Reproducibility in {{Data Science}} and {{Machine Learning}}},
@@ -15537,6 +16951,21 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
file = {/Users/wasmer/Nextcloud/Zotero/Stevens et al_2020_AI for Science Report 2020.pdf;/Users/wasmer/Zotero/storage/Q2KH2Y8A/ai-for-science-report-2020.html}
}
+@online{stierSignificanceAcceleratedDiscovery2023,
+ title = {The {{Significance}} of {{Accelerated Discovery}} of {{Advanced Materials}} to Address {{Societal Challenges}}},
+ author = {Stier, Simon and Kreisbeck, Christoph and Ihssen, Holger and Popp, Matthias Albert and Hauch, Jens and Malek, Kourosh and Reynaud, Marine and Carlsson, Johan and Gold, Lukas and Goumans, Fedor and Todorov, Ilian and Räder, Andreas and Bandesha, Shahbaz Tareq and Wenzel, Wolfgang and Jacques, Philippe and Arcelus, Oier and Garcia-Moreno, Francisco and Friederich, Pascal and Maglione, Mario and Clark, Simon and Laukkanen, Anssi and Cabanas, Montserrat Casas and Carrasco, Javier and Castelli, Ivano Eligio and Stein, Helge Sören and Vegge, Tejs and Nakamae, Sawako and Fabrizio, Monica and Kozdras, Mark},
+ date = {2023-06-07},
+ eprinttype = {Zenodo},
+ doi = {10.5281/zenodo.8012140},
+ url = {https://zenodo.org/records/8012140},
+ urldate = {2024-08-01},
+ abstract = {Societal Challenges demand for Advanced Materials, which in turn promise economical potential. Material Acceleration Platforms (MAPs) will decrease their development time and cost. We comment on implications for science, industry and policy concluding with necessary steps towards establishment of MAPs.},
+ langid = {english},
+ pubstate = {prepublished},
+ keywords = {AML,autonomous research systems,commentary,energy challenge,energy materials,experimental,for introductions,hybrid AI/simulation,integrated models,lab automation,materials,materials acceleration platforms,materials discovery,self-driving lab},
+ file = {/Users/wasmer/Nextcloud/Zotero/Stier et al_2023_The Significance of Accelerated Discovery of Advanced Materials to address.pdf}
+}
+
@article{stocksCompleteSolutionKorringaKohnRostoker1978,
title = {Complete {{Solution}} of the {{Korringa-Kohn-Rostoker Coherent-Potential-Approximation Equations}}: {{Cu-Ni Alloys}}},
shorttitle = {Complete {{Solution}} of the {{Korringa-Kohn-Rostoker Coherent-Potential-Approximation Equations}}},
@@ -15571,7 +17000,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
urldate = {2024-04-18},
abstract = {Despite the successes of machine learning methods in physical sciences, the prediction of the Hamiltonian, and thus the electronic properties, is still unsatisfactory. Based on graph neural network (NN) architecture, we present an extendable NN model to determine the Hamiltonian from ab initio data, with only local atomic structures as inputs. The rotational equivariance of the Hamiltonian is achieved by our complete local coordinates (LCs). The LC information, encoded using a convolutional NN and designed to preserve Hermitian symmetry, is used to map hopping parameters onto local structures. We demonstrate the performance of our model using graphene and SiGe random alloys as examples. We show that our NN model, although trained using small-size systems, can predict the Hamiltonian, as well as electronic properties such as band structures and densities of states for large-size systems within the ab initio accuracy, justifying its extensibility. In combination with the high efficiency of our model, which takes only seconds to get the Hamiltonian of a 1728-atom system, the present work provides a general framework to predict electronic properties efficiently and accurately, which provides new insights into computational physics and will accelerate the research for large-scale materials.},
langid = {english},
- keywords = {\_tablet,AML,attention,CNN,convolution,equivariant,GNN,ML,ML-DFT,ML-ESM,prediction of bandstructure,prediction of DOS,prediction of Hamiltonian matrix,ResNet,skip connection},
+ keywords = {AML,attention,CNN,convolution,equivariant,GNN,ML,ML-DFT,ML-ESM,prediction of bandstructure,prediction of DOS,prediction of Hamiltonian matrix,ResNet,skip connection},
file = {/Users/wasmer/Nextcloud/Zotero/Su et al_2023_Efficient determination of the Hamiltonian and electronic properties using.pdf}
}
@@ -15588,7 +17017,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
urldate = {2023-11-30},
abstract = {According to density functional theory, any chemical property can be inferred from the electron density, making it the most informative attribute of an atomic structure. In this work, we demonstrate the use of established physical methods to obtain important chemical properties from model-predicted electron densities. We introduce graph neural network architectural choices that provide physically relevant and useful electron density predictions. Despite not being trained to predict atomic charges, the model is able to predict atomic charges with an error of an order of magnitude lower than that of a sum of atomic charge densities. Similarly, the model predicts dipole moments with half the error of the sum of the atomic charge densities method. We demonstrate that larger data sets lead to more useful predictions for these tasks. These results pave the way for an alternative path in atomistic machine learning where data-driven approaches and existing physical methods are used in tandem to obtain a variety of chemical properties in an explainable and self-consistent manner.},
keywords = {AML,dipole moments,GNN,hybrid AI/simulation,library,ML,ML-DFT,ML-ESM,OC20,partial charges,prediction of electron density,SchNet,VASP,with-code},
- file = {/Users/wasmer/Nextcloud/Zotero/Sunshine et al_2023_Chemical Properties from Graph Neural Network-Predicted Electron Densities.pdf;/Users/wasmer/Nextcloud/Zotero/Sunshine et al_2023_Chemical Properties from Graph Neural Network-Predicted Electron Densities2.pdf}
+ file = {/Users/wasmer/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Sunshine et al_2023_Chemical Properties from Graph Neural Network-Predicted Electron Densities.pdf}
}
@online{suSVNetWhereEquivariance2022,
@@ -15597,13 +17026,13 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
author = {Su, Zhuo and Welling, Max and Pietikäinen, Matti and Liu, Li},
date = {2022-09-20},
eprint = {2209.05924},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2209.05924},
url = {http://arxiv.org/abs/2209.05924},
urldate = {2023-08-22},
abstract = {Efficiency and robustness are increasingly needed for applications on 3D point clouds, with the ubiquitous use of edge devices in scenarios like autonomous driving and robotics, which often demand real-time and reliable responses. The paper tackles the challenge by designing a general framework to construct 3D learning architectures with SO(3) equivariance and network binarization. However, a naive combination of equivariant networks and binarization either causes sub-optimal computational efficiency or geometric ambiguity. We propose to locate both scalar and vector features in our networks to avoid both cases. Precisely, the presence of scalar features makes the major part of the network binarizable, while vector features serve to retain rich structural information and ensure SO(3) equivariance. The proposed approach can be applied to general backbones like PointNet and DGCNN. Meanwhile, experiments on ModelNet40, ShapeNet, and the real-world dataset ScanObjectNN, demonstrated that the method achieves a great trade-off between efficiency, rotation robustness, and accuracy. The codes are available at https://github.com/zhuoinoulu/svnet.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,todo-tagging},
file = {/Users/wasmer/Nextcloud/Zotero/Su et al_2022_SVNet.pdf;/Users/wasmer/Zotero/storage/YCIPFJTS/2209.html}
}
@@ -15642,7 +17071,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
url = {https://link.aps.org/doi/10.1103/RevModPhys.95.035004},
urldate = {2023-12-14},
abstract = {This review addresses the method of explicit calculations of interatomic exchange interactions of magnetic materials. This involves exchange mechanisms normally referred to as a Heisenberg exchange, a Dzyaloshinskii-Moriya interaction, and an anisotropic symmetric exchange. The connection between microscopic theories of the electronic structure, such as density functional theory and dynamical mean-field theory, and interatomic exchange is examined. The different aspects of extracting information for an effective spin Hamiltonian that involves thousands of atoms, from electronic structure calculations considering significantly fewer atoms (1–50), is highlighted. Examples of exchange interactions of a large group of materials is presented, which involves heavy elements of the 3d period, alloys between transition metals, Heusler compounds, multilayer systems as well as overlayers and adatoms on a substrate, transition metal oxides, 4f elements, magnetic materials in two dimensions, and molecular magnets. Where possible, a comparison to experimental data is made that becomes focused on the magnon dispersion. The influence of relativity is reviewed in a few cases, as is the importance of dynamical correlations. Development to theories that handle out-of-equilibrium conditions is also described here. The review ends with a description of extensions of the theories behind explicit calculations of interatomic exchange to nonmagnetic situations, such as those that describe chemical (charge) order and superconductivity.},
- keywords = {\_tablet,DFT,Dzyaloshinskii–Moriya interaction,educational,exchange interaction,finite-temperature,Green's functions,Heisenberg model,Jij,kinetic exchange,learning material,magnetic interactions,magnetism,physics,quantum magnetism,rec-by-katsumoto,review,SOC,Spin Hamiltonian,spin-dependent,symmetry breaking,transition metals,TRS},
+ keywords = {DFT,Dzyaloshinskii–Moriya interaction,educational,exchange interaction,finite-temperature,Green's functions,Heisenberg model,Jij,kinetic exchange,learning material,magnetic interactions,magnetism,physics,quantum magnetism,rec-by-katsumoto,review,SOC,Spin Hamiltonian,spin-dependent,symmetry breaking,transition metals,TRS},
file = {/Users/wasmer/Nextcloud/Zotero/Szilva et al_2023_Quantitative theory of magnetic interactions in solids.pdf;/Users/wasmer/Zotero/storage/WII6UD6M/RevModPhys.95.html}
}
@@ -15674,6 +17103,60 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
file = {/Users/wasmer/Zotero/storage/IDWYNP9E/www.statlect.com.html}
}
+@article{taborAcceleratingDiscoveryMaterials2018,
+ title = {Accelerating the Discovery of Materials for Clean Energy in the Era of Smart Automation},
+ author = {Tabor, Daniel P. and Roch, Loïc M. and Saikin, Semion K. and Kreisbeck, Christoph and Sheberla, Dennis and Montoya, Joseph H. and Dwaraknath, Shyam and Aykol, Muratahan and Ortiz, Carlos and Tribukait, Hermann and Amador-Bedolla, Carlos and Brabec, Christoph J. and Maruyama, Benji and Persson, Kristin A. and Aspuru-Guzik, Alán},
+ date = {2018-05},
+ journaltitle = {Nature Reviews Materials},
+ shortjournal = {Nat Rev Mater},
+ volume = {3},
+ number = {5},
+ pages = {5--20},
+ publisher = {Nature Publishing Group},
+ issn = {2058-8437},
+ doi = {10.1038/s41578-018-0005-z},
+ url = {https://www.nature.com/articles/s41578-018-0005-z},
+ urldate = {2024-08-01},
+ abstract = {The discovery and development of novel materials in the field of energy are essential to accelerate the transition to a low-carbon economy. Bringing recent technological innovations in automation, robotics and computer science together with current approaches in chemistry, materials synthesis and characterization will act as a catalyst for revolutionizing traditional research and development in both industry and academia. This Perspective provides a vision for an integrated artificial intelligence approach towards autonomous materials discovery, which, in our opinion, will emerge within the next 5 to 10 years. The approach we discuss requires the integration of the following tools, which have already seen substantial development to date: high-throughput virtual screening, automated synthesis planning, automated laboratories and machine learning algorithms. In addition to reducing the time to deployment of new materials by an order of magnitude, this integrated approach is expected to lower the cost associated with the initial discovery. Thus, the price of the final products (for example, solar panels, batteries and electric vehicles) will also decrease. This in turn will enable industries and governments to meet more ambitious targets in terms of reducing greenhouse gas emissions at a faster pace.},
+ langid = {english},
+ keywords = {autonomous research systems,for introductions,lab automation,materials acceleration platforms,materials discovery,self-driving lab},
+ file = {/Users/wasmer/Nextcloud/Zotero/Tabor et al_2018_Accelerating the discovery of materials for clean energy in the era of smart.pdf}
+}
+
+@article{tahmasbiMachineLearningdrivenStructure2024,
+ title = {Machine Learning-Driven Structure Prediction for Iron Hydrides},
+ author = {Tahmasbi, Hossein and Ramakrishna, Kushal and Lokamani, Mani and Cangi, Attila},
+ date = {2024-03-21},
+ journaltitle = {Physical Review Materials},
+ shortjournal = {Phys. Rev. Mater.},
+ volume = {8},
+ number = {3},
+ pages = {033803},
+ publisher = {American Physical Society},
+ doi = {10.1103/PhysRevMaterials.8.033803},
+ url = {https://link.aps.org/doi/10.1103/PhysRevMaterials.8.033803},
+ urldate = {2024-06-07},
+ abstract = {We created a computational workflow to analyze the potential energy surface (PES) of materials using machine-learned interatomic potentials in conjunction with the minima hopping algorithm. We demonstrate this method by producing a versatile machine-learned interatomic potential for iron hydride via a neural network using an iterative training process to explore its energy landscape under different pressures. To evaluate the accuracy and comprehend the intricacies of the PES, we conducted comprehensive crystal structure predictions using our neural network-based potential paired with the minima hopping approach. The predictions spanned pressures ranging from ambient to 100 GPa. Our results reproduce the experimentally verified global minimum structures such as dhcp, hcp, and fcc, corroborating previous findings. Furthermore, our in-depth exploration of the iron hydride PES at different pressures has revealed complex alterations and stacking faults in these phases, leading to the identification of several different low-enthalpy structures. This investigation has not only confirmed the presence of regions of established FeH configurations but has also highlighted the efficacy of using data-driven, extensive structure prediction methods to uncover the multifaceted PES of materials.},
+ keywords = {/unread,active learning,alloys,AML,binary systems,CASUS,FLAME,iron,ML,MLP,structure prediction},
+ file = {/Users/wasmer/Nextcloud/Zotero/Tahmasbi et al_2024_Machine learning-driven structure prediction for iron hydrides.pdf;/Users/wasmer/Zotero/storage/VZE84HFX/PhysRevMaterials.8.html}
+}
+
+@article{takahashiFullyAutonomousMaterials2023,
+ title = {Fully Autonomous Materials Screening Methodology Combining First-Principles Calculations, Machine Learning and High-Performance Computing System},
+ author = {Takahashi, Akira and Terayama, Kei and Kumagai, Yu and Tamura, Ryo and Oba, Fumiyasu},
+ date = {2023-12-31},
+ journaltitle = {Science and Technology of Advanced Materials: Methods},
+ publisher = {Taylor \& Francis},
+ issn = {2766-0400},
+ doi = {10.1080/27660400.2023.2261834},
+ url = {https://www.tandfonline.com/doi/abs/10.1080/27660400.2023.2261834%4010.1080/tfocoll.2023.0.issue-MLAA},
+ urldate = {2024-05-24},
+ abstract = {Materials screening by high-throughput first-principles calculations is a powerful tool for exploring novel materials with preferable properties. Machine learning techniques are expected to acceler...},
+ langid = {english},
+ keywords = {/unread,active learning,Bayesian methods,black-box optimization,compositional descriptors,descriptors,DFT,HTC,kernel methods,materials discovery,materials screening,MatMiner,ML,random forest,with-code,workflows},
+ file = {/Users/wasmer/Nextcloud/Zotero/Takahashi et al_2023_Fully autonomous materials screening methodology combining first-principles.pdf;/Users/wasmer/Zotero/storage/5785PJ42/tfocoll.2023.0.html}
+}
+
@article{takamotoTeaNetUniversalNeural2022,
title = {{{TeaNet}}: {{Universal}} Neural Network Interatomic Potential Inspired by Iterative Electronic Relaxations},
shorttitle = {{{TeaNet}}},
@@ -15689,7 +17172,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
urldate = {2023-06-30},
abstract = {A universal interatomic potential for an arbitrary set of chemical elements is urgently needed in computational materials science. Graph convolution neural network (GCN) has rich expressive power, but previously was mainly employed to transport scalars and vectors, not rank ≥2 tensors. As classic interatomic potentials were inspired by tight-binding electronic relaxation framework, we want to represent this iterative propagation of rank ≥2 tensor information by GCN. Here we propose an architecture called the tensor embedded atom network (TeaNet) where angular interaction is translated into graph convolution through the incorporation of Euclidean tensors, vectors and scalars. By applying the residual network (ResNet) architecture and training with recurrent GCN weights initialization, a much deeper (16 layers) GCN was constructed, whose flow is similar to an iterative electronic relaxation. Our training dataset is generated by density functional theory calculation of mostly chemically and structurally randomized configurations. We demonstrate that arbitrary structures and reactions involving the first 18 elements on the periodic table (H to Ar) can be realized satisfactorily by TeaNet, including C–H molecular structures, metals, amorphous SiO2, and water, showing surprisingly good performance (energy mean absolute error 19 meV/atom) and robustness for arbitrary chemistries involving elements from H to Ar.},
langid = {english},
- keywords = {/unread,AML,GNN,ML,MLP,TeaNet,tensorial target,universal potential},
+ keywords = {AML,GNN,ML,MLP,TeaNet,tensorial target,universal potential,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Takamoto et al_2022_TeaNet.pdf;/Users/wasmer/Zotero/storage/BF944L4Z/S0927025622000799.html}
}
@@ -15710,8 +17193,8 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
abstract = {Computational material discovery is under intense study owing to its ability to explore the vast space of chemical systems. Neural network potentials (NNPs) have been shown to be particularly effective in conducting atomistic simulations for such purposes. However, existing NNPs are generally designed for narrow target materials, making them unsuitable for broader applications in material discovery. Here we report a development of universal NNP called PreFerred Potential (PFP), which is able to handle any combination of 45 elements. Particular emphasis is placed on the datasets, which include a diverse set of virtual structures used to attain the universality. We demonstrated the applicability of PFP in selected domains: lithium diffusion in LiFeSO4F, molecular adsorption in metal-organic frameworks, an order–disorder transition of Cu-Au alloys, and material discovery for a Fischer–Tropsch catalyst. They showcase the power of PFP, and this technology provides a highly useful tool for material discovery.},
issue = {1},
langid = {english},
- keywords = {/unread,AML,ML,MLP,PreFerred Potential,universal potential},
- file = {/Users/wasmer/Nextcloud/Zotero/Takamoto et al_2022_Towards universal neural network potential for material discovery applicable to.pdf}
+ keywords = {AML,closed-source,DimeNet++,disordered,Ferromagnetism,ML,MLP,OC20,PBE,PreFerred Potential,proprietary,SchNet,speedup,spin-polarized,TeaNet,tensorial target,universal potential,VASP,with-code,with-data},
+ file = {/Users/wasmer/Nextcloud/Zotero/Takamoto et al_2022_Towards universal neural network potential for material discovery applicable to.pdf;/Users/wasmer/Nextcloud/Zotero/Takamoto et al_2022_Towards universal neural network potential for material discovery applicable to2.pdf}
}
@article{takamotoUniversalNeuralNetwork2023,
@@ -15728,7 +17211,7 @@ Subject\_term\_id: condensed-matter-physics;electronic-structure;materials-scien
url = {https://www.sciencedirect.com/science/article/pii/S2352847823000072},
urldate = {2023-06-30},
langid = {english},
- keywords = {AML,GNN,ML,MLP,prediction of electron density,PreFerred Potential,TeaNet,tensorial target,universal potential},
+ keywords = {AML,closed-source,commercial,GNN,ML,MLP,Peer review,prediction of electron density,PreFerred Potential,TeaNet,tensorial target,universal potential,with-review},
file = {/Users/wasmer/Nextcloud/Zotero/Takamoto et al_2023_Towards universal neural network interatomic potential.pdf;/Users/wasmer/Zotero/storage/XEJZUGIH/S2352847823000072.html}
}
@@ -15763,7 +17246,7 @@ Subject\_term\_id: databases;materials-science},
author = {Talirz, Leopold and Ghiringhelli, Luca M. and Smit, Berend},
date = {2021-08-27},
eprint = {2108.12350},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2108.12350},
urldate = {2021-09-11},
@@ -15796,13 +17279,13 @@ Subject\_term\_id: databases;materials-science},
author = {Tang, Zechen and Li, He and Lin, Peize and Gong, Xiaoxun and Jin, Gan and He, Lixin and Jiang, Hong and Ren, Xinguo and Duan, Wenhui and Xu, Yong},
date = {2023-02-16},
eprint = {2302.08221},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2302.08221},
url = {http://arxiv.org/abs/2302.08221},
urldate = {2024-04-18},
abstract = {Hybrid density functional calculation is indispensable to accurate description of electronic structure, whereas the formidable computational cost restricts its broad application. Here we develop a deep equivariant neural network method (named DeepH-hybrid) to learn the hybrid-functional Hamiltonian from self-consistent field calculations of small structures, and apply the trained neural networks for efficient electronic-structure calculation by passing the self-consistent iterations. The method is systematically checked to show high efficiency and accuracy, making the study of large-scale materials with hybrid-functional accuracy feasible. As an important application, the DeepH-hybrid method is applied to study large-supercell Moir\textbackslash '\{e\} twisted materials, offering the first case study on how the inclusion of exact exchange affects flat bands in the magic-angle twisted bilayer graphene.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,DeepH,hybrid DFT,ML,ML-DFT,ML-ESM,no-code,prediction of Hamiltonian matrix,transfer learning},
file = {/Users/wasmer/Nextcloud/Zotero/Tang et al_2023_Efficient hybrid density functional calculation by deep learning.pdf;/Users/wasmer/Zotero/storage/E995LV4K/2302.html}
}
@@ -15853,13 +17336,13 @@ Ying-Wai Li\\
author = {Teufel, Jonas and Torresi, Luca and Reiser, Patrick and Friederich, Pascal},
date = {2022-11-23},
eprint = {2211.13236},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2211.13236},
url = {http://arxiv.org/abs/2211.13236},
urldate = {2023-03-02},
abstract = {Explainable artificial intelligence (XAI) methods are expected to improve trust during human-AI interactions, provide tools for model analysis and extend human understanding of complex problems. Explanation-supervised training allows to improve explanation quality by training self-explaining XAI models on ground truth or human-generated explanations. However, existing explanation methods have limited expressiveness and interoperability due to the fact that only single explanations in form of node and edge importance are generated. To that end we propose the novel multi-explanation graph attention network (MEGAN). Our fully differentiable, attention-based model features multiple explanation channels, which can be chosen independently of the task specifications. We first validate our model on a synthetic graph regression dataset. We show that for the special single explanation case, our model significantly outperforms existing post-hoc and explanation-supervised baseline methods. Furthermore, we demonstrate significant advantages when using two explanations, both in quantitative explanation measures as well as in human interpretability. Finally, we demonstrate our model's capabilities on multiple real-world datasets. We find that our model produces sparse high-fidelity explanations consistent with human intuition about those tasks and at the same time matches state-of-the-art graph neural networks in predictive performance, indicating that explanations and accuracy are not necessarily a trade-off.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,GAT,GNN,materials,ML,XAI},
file = {/Users/wasmer/Nextcloud/Zotero/Teufel et al_2022_MEGAN.pdf;/Users/wasmer/Zotero/storage/4ZA3I5AT/2211.html}
}
@@ -15931,7 +17414,7 @@ Ying-Wai Li\\
volume = {85},
number = {23},
doi = {10.1103/PhysRevB.85.235103},
- keywords = {\_tablet,juKKR,KKR,KKRnano,PGI-1/IAS-1},
+ keywords = {juKKR,KKR,KKRnano,PGI-1/IAS-1},
file = {/Users/wasmer/Nextcloud/Zotero/Thiess_2012_Massively parallel density functional calculations for thousands of atoms.pdf;/Users/wasmer/Zotero/storage/PM97ULPL/PhysRevB.85.html}
}
@@ -15955,13 +17438,13 @@ Ying-Wai Li\\
author = {Thomas, Nathaniel and Smidt, Tess and Kearnes, Steven and Yang, Lusann and Li, Li and Kohlhoff, Kai and Riley, Patrick},
date = {2018-05-18},
eprint = {1802.08219},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.1802.08219},
url = {http://arxiv.org/abs/1802.08219},
urldate = {2023-06-30},
abstract = {We introduce tensor field neural networks, which are locally equivariant to 3D rotations, translations, and permutations of points at every layer. 3D rotation equivariance removes the need for data augmentation to identify features in arbitrary orientations. Our network uses filters built from spherical harmonics; due to the mathematical consequences of this filter choice, each layer accepts as input (and guarantees as output) scalars, vectors, and higher-order tensors, in the geometric sense of these terms. We demonstrate the capabilities of tensor field networks with tasks in geometry, physics, and chemistry.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,classical mechanics,continuous convolution,convolution,e3nn,equivariant,General ML,geometric tensors,ML,molecules,point cloud data,shape classification,tensor field,tensorial target,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Thomas et al_2018_Tensor field networks.pdf;/Users/wasmer/Zotero/storage/2U9EVPUR/1802.html}
}
@@ -16008,7 +17491,7 @@ Ying-Wai Li\\
author = {Togo, Atsushi and Tanaka, Isao},
date = {2018-08-05},
eprint = {1808.01590},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
publisher = {arXiv},
doi = {10.48550/arXiv.1808.01590},
@@ -16078,7 +17561,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
abstract = {The importance of global band topology is unequivocally recognized in condensed matter physics, and new states of matter, such as topological insulators, have been discovered. Owing to their bulk band topology, 3D topological insulators possess a massless Dirac dispersion with spin–momentum locking at the surface. Although 3D topological insulators were originally proposed in time-reversal invariant systems, the onset of a spontaneous magnetization or, equivalently, a broken time-reversal symmetry leads to the formation of an exchange gap in the Dirac band dispersion. In such magnetic topological insulators, tuning of the Fermi level in the exchange gap results in the emergence of a quantum Hall effect at zero magnetic field, that is, of a quantum anomalous Hall effect. Here, we review the basic concepts of magnetic topological insulators and their experimental realization, together with the discovery and verification of their emergent properties. In particular, we discuss how the development of tailored materials through heterostructure engineering has made it possible to access the quantum anomalous Hall effect, the topological magnetoelectric effect, the physics related to the chiral edge states that appear in these materials and various spintronic phenomena. Further theoretical and experimental research on magnetic topological insulators will provide fertile ground for the development of new concepts for next-generation electronic devices for applications such as spintronics with low energy consumption, dissipationless topological electronics and topological quantum computation.},
issue = {2},
langid = {english},
- keywords = {\_tablet,breaking of TRS,Chern number,Hall effect,Hall QAHE,Hall QHE,heterostructures,magnetic doping,magnetic TIs,Majorana,materials,physics,review,spin-momentum locking,Spintronics,TKNN,topological insulator,transition metals,TRS},
+ keywords = {breaking of TRS,Chern number,for introductions,Hall effect,Hall QAHE,Hall QHE,heterostructures,magnetic doping,magnetic TIs,Majorana,materials,physics,review,spin-momentum locking,Spintronics,TKNN,topological insulator,transition metals,TRS},
file = {/Users/wasmer/Nextcloud/Zotero/Tokura et al_2019_Magnetic topological insulators.pdf}
}
@@ -16142,13 +17625,13 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {Townshend, Raphael J. L. and Vögele, Martin and Suriana, Patricia and Derry, Alexander and Powers, Alexander and Laloudakis, Yianni and Balachandar, Sidhika and Jing, Bowen and Anderson, Brandon and Eismann, Stephan and Kondor, Risi and Altman, Russ B. and Dror, Ron O.},
date = {2022-01-15},
eprint = {2012.04035},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, q-bio},
doi = {10.48550/arXiv.2012.04035},
url = {http://arxiv.org/abs/2012.04035},
urldate = {2023-10-05},
abstract = {Computational methods that operate on three-dimensional molecular structure have the potential to solve important questions in biology and chemistry. In particular, deep neural networks have gained significant attention, but their widespread adoption in the biomolecular domain has been limited by a lack of either systematic performance benchmarks or a unified toolkit for interacting with molecular data. To address this, we present ATOM3D, a collection of both novel and existing benchmark datasets spanning several key classes of biomolecules. We implement several classes of three-dimensional molecular learning methods for each of these tasks and show that they consistently improve performance relative to methods based on one- and two-dimensional representations. The specific choice of architecture proves to be critical for performance, with three-dimensional convolutional networks excelling at tasks involving complex geometries, graph networks performing well on systems requiring detailed positional information, and the more recently developed equivariant networks showing significant promise. Our results indicate that many molecular problems stand to gain from three-dimensional molecular learning, and that there is potential for improvement on many tasks which remain underexplored. To lower the barrier to entry and facilitate further developments in the field, we also provide a comprehensive suite of tools for dataset processing, model training, and evaluation in our open-source atom3d Python package. All datasets are available for download from https://www.atom3d.ai .},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,benchmark dataset,benchmarking,biomolecules,chemistry,Database,library,ML,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Townshend et al_2022_ATOM3D2.pdf;/Users/wasmer/Zotero/storage/2D2AMNSF/2012.html}
}
@@ -16158,14 +17641,14 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {Tran, Kevin and Neiswanger, Willie and Yoon, Junwoong and Zhang, Qingyang and Xing, Eric and Ulissi, Zachary W.},
date = {2020-02-20},
eprint = {1912.10066},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.1912.10066},
url = {http://arxiv.org/abs/1912.10066},
urldate = {2023-04-11},
abstract = {Data science and informatics tools have been proliferating recently within the computational materials science and catalysis fields. This proliferation has spurned the creation of various frameworks for automated materials screening, discovery, and design. Underpinning these frameworks are surrogate models with uncertainty estimates on their predictions. These uncertainty estimates are instrumental for determining which materials to screen next, but the computational catalysis field does not yet have a standard procedure for judging the quality of such uncertainty estimates. Here we present a suite of figures and performance metrics derived from the machine learning community that can be used to judge the quality of such uncertainty estimates. This suite probes the accuracy, calibration, and sharpness of a model quantitatively. We then show a case study where we judge various methods for predicting density-functional-theory-calculated adsorption energies. Of the methods studied here, we find that the best performer is a model where a convolutional neural network is used to supply features to a Gaussian process regressor, which then makes predictions of adsorption energies along with corresponding uncertainty estimates.},
- pubstate = {preprint},
- keywords = {active learning,AML,Bayesian methods,Gaussian process,materials,ML,regression,uncertainty quantification},
+ pubstate = {prepublished},
+ keywords = {active learning,AML,Bayesian methods,Gaussian process,materials,ML,neural network,regression,uncertainty quantification},
file = {/Users/wasmer/Nextcloud/Zotero/Tran et al_2020_Methods for comparing uncertainty quantifications for material property.pdf;/Users/wasmer/Zotero/storage/6RLGREQU/1912.html}
}
@@ -16178,7 +17661,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
volume = {13},
number = {5},
eprint = {2206.08917},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
pages = {3066--3084},
issn = {2155-5435, 2155-5435},
@@ -16206,7 +17689,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
abstract = {The overwhelming majority of scientific knowledge is published as text, which is difficult to analyse by either traditional statistical analysis or modern machine learning methods. By contrast, the main source of machine-interpretable data for the materials research community has come from structured property databases1,2, which encompass only a small fraction of the knowledge present in the research literature. Beyond property values, publications contain valuable knowledge regarding the connections and relationships between data items as interpreted by the authors. To improve the identification and use of this knowledge, several studies have focused on the retrieval of information from scientific literature using supervised natural language processing3–10, which requires large hand-labelled datasets for training. Here we show that materials science knowledge present in the published literature can be efficiently encoded as information-dense word embeddings11–13 (vector representations of words) without human labelling or supervision. Without any explicit insertion of chemical knowledge, these embeddings capture complex materials science concepts such as the underlying structure of the periodic table and structure–property relationships in materials. Furthermore, we demonstrate that an unsupervised method can recommend materials for functional applications several years before their discovery. This suggests that latent knowledge regarding future discoveries is to a large extent embedded in past publications. Our findings highlight the possibility of extracting knowledge and relationships from the massive body of scientific literature in a collective manner, and point towards a generalized approach to the mining of scientific literature.},
issue = {7763},
langid = {english},
- keywords = {/unread,AML,descriptors,embedding,literature analysis,Mat2Vec,materials,ML,nlp,unsupervised learning,with-code,Word2Vec},
+ keywords = {/unread,AML,descriptors,embedding,language models,literature analysis,Mat2Vec,materials,ML,nlp,unsupervised learning,with-code,Word2Vec},
file = {/Users/wasmer/Zotero/storage/NEI3YJWG/Tshitoyan et al. - 2019 - Unsupervised word embeddings capture latent knowle.pdf}
}
@@ -16318,7 +17801,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
url = {https://proceedings.neurips.cc/paper/2021/hash/78f1893678afbeaa90b1fa01b9cfb860-Abstract.html},
urldate = {2022-08-21},
abstract = {Machine learning has enabled the prediction of quantum chemical properties with high accuracy and efficiency, allowing to bypass computationally costly ab initio calculations. Instead of training on a fixed set of properties, more recent approaches attempt to learn the electronic wavefunction (or density) as a central quantity of atomistic systems, from which all other observables can be derived. This is complicated by the fact that wavefunctions transform non-trivially under molecular rotations, which makes them a challenging prediction target. To solve this issue, we introduce general SE(3)-equivariant operations and building blocks for constructing deep learning architectures for geometric point cloud data and apply them to reconstruct wavefunctions of atomistic systems with unprecedented accuracy. Our model achieves speedups of over three orders of magnitude compared to ab initio methods and reduces prediction errors by up to two orders of magnitude compared to the previous state-of-the-art. This accuracy makes it possible to derive properties such as energies and forces directly from the wavefunction in an end-to-end manner. We demonstrate the potential of our approach in a transfer learning application, where a model trained on low accuracy reference wavefunctions implicitly learns to correct for electronic many-body interactions from observables computed at a higher level of theory. Such machine-learned wavefunction surrogates pave the way towards novel semi-empirical methods, offering resolution at an electronic level while drastically decreasing computational cost. Additionally, the predicted wavefunctions can serve as initial guess in conventional ab initio methods, decreasing the number of iterations required to arrive at a converged solution, thus leading to significant speedups without any loss of accuracy or robustness. While we focus on physics applications in this contribution, the proposed equivariant framework for deep learning on point clouds is promising also beyond, say, in computer vision or graphics.},
- keywords = {\_tablet,EGNN,equivariant,initial guess,ML-ESM,original publication,PhiSNet,prediction of electron density,prediction of wavefunction,SchNOrb,with-review},
+ keywords = {EGNN,equivariant,initial guess,ML-ESM,original publication,PhiSNet,prediction of electron density,prediction of wavefunction,SchNOrb,with-review},
file = {/Users/wasmer/Nextcloud/Zotero/Unke et al_2021_SE(3)-equivariant prediction of molecular wavefunctions and electronic densities.pdf}
}
@@ -16339,7 +17822,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
abstract = {Atomistic modeling of chemically reactive systems has so far relied on either expensive ab initio methods or bond-order force fields requiring arduous parametrization. Here, we describe a Bayesian active learning framework for autonomous “on-the-fly†training of fast and accurate reactive many-body force fields during molecular dynamics simulations. At each time-step, predictive uncertainties of a sparse Gaussian process are evaluated to automatically determine whether additional ab initio training data are needed. We introduce a general method for mapping trained kernel models onto equivalent polynomial models whose prediction cost is much lower and independent of the training set size. As a demonstration, we perform direct two-phase simulations of heterogeneous H2 turnover on the Pt(111) catalyst surface at chemical accuracy. The model trains itself in three days and performs at twice the speed of a ReaxFF model, while maintaining much higher fidelity to DFT and excellent agreement with experiment.},
issue = {1},
langid = {english},
- keywords = {/unread,\_tablet,active learning,active learning online,AML,Bayesian methods,FLARE,Gaussian process,GPR,iterative learning,library,MD,ML,MLP,uncertainty quantification,with-code},
+ keywords = {active learning,active learning online,AML,Bayesian methods,FLARE,Gaussian process,GPR,iterative learning,library,MD,ML,MLP,uncertainty quantification,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Vandermause et al_2022_Active learning of reactive Bayesian force fields applied to heterogeneous.pdf}
}
@@ -16360,7 +17843,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
abstract = {Data-driven interatomic potentials have emerged as a powerful tool for approximating ab initio potential energy surfaces. The most time-consuming step in creating these interatomic potentials is typically the generation of a suitable training database. To aid this process hyperactive learning (HAL), an accelerated active learning scheme, is presented as a method for rapid automated training database assembly. HAL adds a biasing term to a physically motivated sampler (e.g. molecular dynamics) driving atomic structures towards uncertainty in turn generating unseen or valuable training configurations. The proposed HAL framework is used to develop atomic cluster expansion (ACE) interatomic potentials for the AlSi10 alloy and polyethylene glycol (PEG) polymer starting from roughly a dozen initial configurations. The HAL generated ACE potentials are shown to be able to determine macroscopic properties, such as melting temperature and density, with close to experimental accuracy.},
issue = {1},
langid = {english},
- keywords = {\_tablet,ACE,active learning,alloys,AML,Bayesian methods,Bayesian optimization,Bayesian regression,binary systems,database generation,HAL,HAL-MD,iterative learning,iterative learning scheme,library,MD,MD17,ML,MLP,molecules,uncertainty quantification,with-code},
+ keywords = {ACE,active learning,alloys,AML,Bayesian methods,Bayesian optimization,Bayesian regression,binary systems,database generation,HAL,HAL-MD,iterative learning,iterative learning scheme,library,MD,MD17,ML,MLP,molecules,uncertainty quantification,with-code},
file = {/Users/wasmer/Zotero/storage/S3FJEDUC/van der Oord et al_2023_Hyperactive learning for data-driven interatomic potentials.pdf}
}
@@ -16369,13 +17852,13 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {family=Oord, given=Cas, prefix=van der, useprefix=true and Sachs, Matthias and Kovács, Dávid Péter and Ortner, Christoph and Csányi, Gábor},
date = {2022-11-07},
eprint = {2210.04225},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.2210.04225},
url = {http://arxiv.org/abs/2210.04225},
urldate = {2023-02-05},
abstract = {Data-driven interatomic potentials have emerged as a powerful class of surrogate models for \{\textbackslash it ab initio\} potential energy surfaces that are able to reliably predict macroscopic properties with experimental accuracy. In generating accurate and transferable potentials the most time-consuming and arguably most important task is generating the training set, which still requires significant expert user input. To accelerate this process, this work presents \textbackslash text\{\textbackslash it hyperactive learning\} (HAL), a framework for formulating an accelerated sampling algorithm specifically for the task of training database generation. The key idea is to start from a physically motivated sampler (e.g., molecular dynamics) and add a biasing term that drives the system towards high uncertainty and thus to unseen training configurations. Building on this framework, general protocols for building training databases for alloys and polymers leveraging the HAL framework will be presented. For alloys, ACE potentials for AlSi10 are created by fitting to a minimal HAL-generated database containing 88 configurations (32 atoms each) with fast evaluation times of {$<$}100 microsecond/atom/cpu-core. These potentials are demonstrated to predict the melting temperature with excellent accuracy. For polymers, a HAL database is built using ACE, able to determine the density of a long polyethylene glycol (PEG) polymer formed of 200 monomer units with experimental accuracy by only fitting to small isolated PEG polymers with sizes ranging from 2 to 32.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ACE,active learning,alloys,AML,Bayesian methods,Bayesian optimization,Bayesian regression,binary systems,database generation,HAL,HAL-MD,iterative learning,iterative learning scheme,library,MD,MD17,ML,MLP,molecules,uncertainty quantification,with-code},
file = {/Users/wasmer/Zotero/storage/4S2GHGVG/van der Oord et al. - 2022 - Hyperactive Learning (HAL) for Data-Driven Interat.pdf;/Users/wasmer/Zotero/storage/YJBLUYLE/2210.html}
}
@@ -16420,7 +17903,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
volume = {15},
number = {2},
eprint = {1407.7722},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
pages = {49--60},
issn = {1931-0145, 1931-0153},
doi = {10.1145/2641190.2641198},
@@ -16465,7 +17948,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
urldate = {2022-09-20},
abstract = {Andreas Berger CICnanoGUNE BRTA Following the success and relevance of the 2014 and 2017 Magnetism Roadmap articles, this 2020 Magnetism Roadmap edition takes yet another timely look at newly relevant and highly active areas in magnetism research. The overall layout of this article is unchanged, given that it has proved the most appropriate way to convey the most relevant aspects of today’s magnetism research in a wide variety of sub-fields to a broad readership. A different group of experts has again been selected for this article, representing both the breadth of new research areas, and the desire to incorporate different voices and viewpoints. The latter is especially relevant for thistype of article, in which one’s field of expertise has to be accommodated on two printed pages only, so that personal selection preferences are naturally rather more visible than in other types of articles. Most importantly, the very relevant advances in the field of magnetism research in recent years make the publication of yet another Magnetism Roadmap a very sensible and timely endeavour, allowing its authors and readers to take another broad-based, but concise look at the most significant developments in magnetism, their precise status, their challenges, and their anticipated future developments. While many of the contributions in this 2020 Magnetism Roadmap edition have significant associations with different aspects of magnetism, the general layout can nonetheless be classified in terms of three main themes: (i) phenomena, (ii) materials and characterization, and (iii) applications and devices. While these categories are unsurprisingly rather similar to the 2017 Roadmap, the order is different, in that the 2020 Roadmap considers phenomena first, even if their occurrences are naturally very difficult to separate from the materials exhibiting such phenomena. Nonetheless, the specifically selected topics seemed to be best displayed in the order presented here, in particular, because many of the phenomena or geometries discussed in (i) can be found or designed into a large variety of materials, so that the progression of the article embarks from more general concepts to more specific classes of materials in the selected order. Given that applications and devices are based on both phenomena and materials, it seemed most appropriate to close the article with the application and devices section (iii) once again. The 2020 Magnetism Roadmap article contains 14 sections, all of which were written by individual authors and experts, specifically addressing a subject in terms of its status, advances, challenges and perspectives in just two pages. Evidently, this two-page format limits the depth to which each subject can be described. Nonetheless, the most relevant and key aspects of each field are touched upon, which enables the Roadmap as whole to give its readership an initial overview of and outlook into a wide variety of topics and fields in a fairly condensed format. Correspondingly, the Roadmap pursues the goal of giving each reader a brief reference frame of relevant and current topics in modern applied magnetism research, even if not all sub-fields can be represented here. The first block of this 2020 Magnetism Roadmap, which is focussed on (i) phenomena, contains five contributions, which address the areas of interfacial Dzyaloshinskii–Moriya interactions, and two-dimensional and curvilinear magnetism, as well as spin-orbit torque phenomena and all optical magnetization reversal. All of these contributions describe cutting edge aspects of rather fundamental physical processes and properties, associated with new and improved magnetic materials’ properties, together with potential developments in terms of future devices and technology. As such, they form part of a widening magnetism ‘phenomena reservoir’ for utilization in applied magnetism and related device technology. The final block (iii) of this article focuses on such applications and device-related fields in four contributions relating to currently active areas of research, which are of course utilizing magnetic phenomena to enable specific functions. These contributions highlight the role of magnetism or spintronics in the field of neuromorphic and reservoir computing, terahertz technology, and domain wall-based logic. One aspect common to all of these application-related contributions is that they are not yet being utilized in commercially available technology; it is currently still an open question, whether or not such technological applications will be magnetism-based at all in the future, or if other types of materials and phenomena will yet outperform magnetism. This last point is actually a very good indication of the vibrancy of applied magnetism research today, given that it demonstrates that magnetism research is able to venture into novel application fields, based upon its portfolio of phenomena, effects and materials. This materials portfolio in particular defines the central block (ii) of this article, with its five contributions interconnecting phenomena with devices, for which materials and the characterization of their properties is the decisive discriminator between purely academically interesting aspects and the true viability of real-life devices, because only available materials and their associated fabrication and characterization methods permit reliable technological implementation. These five contributions specifically address magnetic films and multiferroic heterostructures for the purpose of spin electronic utilization, multi-scale materials modelling, and magnetic materials design based upon machine-learning, as well as materials characterization via polarized neutron measurements. As such, these contributions illustrate the balanced relevance of research into experimental and modelling magnetic materials, as well the importance of sophisticated characterization methods that allow for an ever-more refined understanding of materials. As a combined and integrated article, this 2020 Magnetism Roadmap is intended to be a reference point for current, novel and emerging research directions in modern magnetism, just as its 2014 and 2017 predecessors have been in previous years.},
langid = {english},
- keywords = {magnetism,physics,review,roadmap},
+ keywords = {\_tablet,magnetism,physics,review,roadmap},
file = {/Users/wasmer/Nextcloud/Zotero/Vedmedenko et al_2020_The 2020 magnetism roadmap.pdf}
}
@@ -16503,13 +17986,13 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {Venugopal, Vineeth and Pai, Sumit and Olivetti, Elsa},
date = {2022-10-31},
eprint = {2210.17340},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2210.17340},
url = {http://arxiv.org/abs/2210.17340},
urldate = {2023-11-05},
abstract = {This paper introduces MatKG, a novel graph database of key concepts in material science spanning the traditional material-structure-property-processing paradigm. MatKG is autonomously generated through transformer-based, large language models and generates pseudo ontological schema through statistical co-occurrence mapping. At present, MatKG contains over 2 million unique relationship triples derived from 80,000 entities. This allows the curated analysis, querying, and visualization of materials knowledge at unique resolution and scale. Further, Knowledge Graph Embedding models are used to learn embedding representations of nodes in the graph which are used for downstream tasks such as link prediction and entity disambiguation. MatKG allows the rapid dissemination and assimilation of data when used as a knowledge base, while enabling the discovery of new relations when trained as an embedding model.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,dataset,graph embedding,knowledge graph,knowledge graph embedding,materials,ML,representation learning},
file = {/Users/wasmer/Nextcloud/Zotero/Venugopal et al_2022_MatKG.pdf;/Users/wasmer/Zotero/storage/GBSDFV8K/2210.html}
}
@@ -16543,6 +18026,18 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
file = {/Users/wasmer/Nextcloud/Zotero/Villar et al_2021_Scalars are universal.pdf}
}
+@unpublished{vindigniIntroductionMagnetism2023,
+ title = {Introduction to {{Magnetism}}},
+ author = {Vindigni, Alessandro},
+ date = {2023},
+ url = {https://vindigni.ch/intro-mag-hs23/},
+ urldate = {2024-06-05},
+ abstract = {This the learning material for the course Introduction to Magnetism HS2022 at ETH Zurich.},
+ howpublished = {Lecture script},
+ keywords = {/unread,condensed matter,Dzyaloshinskii–Moriya interaction,educational,exchange interaction,finite-temperature,Heisenberg model,Ising,itinerant magnetism,Jij,learning material,lecture notes,magnetic order,magnetism,online learning,physics,rare earths,RKKY interaction,spin models,transition metals},
+ file = {/Users/wasmer/Nextcloud/Zotero/Vindigni_Introduction to Magnetism HS22.pdf;/Users/wasmer/Zotero/storage/8L3BUVYE/intro-mag-hs23.html}
+}
+
@article{vojvodicExploringLimitsLowpressure2014,
title = {Exploring the Limits: {{A}} Low-Pressure, Low-Temperature {{Haber}}–{{Bosch}} Process},
shorttitle = {Exploring the Limits},
@@ -16567,13 +18062,13 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {family=Glehn, given=Ingrid, prefix=von, useprefix=true and Spencer, James S. and Pfau, David},
date = {2022-11-24},
eprint = {2211.13672},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2211.13672},
url = {http://arxiv.org/abs/2211.13672},
urldate = {2023-04-17},
abstract = {We present a novel neural network architecture using self-attention, the Wavefunction Transformer (Psiformer), which can be used as an approximation (or Ansatz) for solving the many-electron Schr\textbackslash "odinger equation, the fundamental equation for quantum chemistry and material science. This equation can be solved from first principles, requiring no external training data. In recent years, deep neural networks like the FermiNet and PauliNet have been used to significantly improve the accuracy of these first-principle calculations, but they lack an attention-like mechanism for gating interactions between electrons. Here we show that the Psiformer can be used as a drop-in replacement for these other neural networks, often dramatically improving the accuracy of the calculations. On larger molecules especially, the ground state energy can be improved by dozens of kcal/mol, a qualitative leap over previous methods. This demonstrates that self-attention networks can learn complex quantum mechanical correlations between electrons, and are a promising route to reaching unprecedented accuracy in chemical calculations on larger systems.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,attention,CCSD(T),DeepMind,FermiNet,JAX,Many-body theory,ML,ML-ESM,ML-QM,ML-QMBP,molecules,PauliNet,prediction of wavefunction,Psiformer,QMC,Quantum chemistry,self-attention,Slater-Jastrow,transfer learning,transformer,VMC},
file = {/Users/wasmer/Nextcloud/Zotero/von Glehn et al_2022_A Self-Attention Ansatz for Ab-initio Quantum Chemistry.pdf;/Users/wasmer/Zotero/storage/DJD69694/2211.html}
}
@@ -16607,7 +18102,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
urldate = {2024-01-01},
abstract = {Transferability, especially in the context of model generalization, is a paradigm of all scientific disciplines. However, the rapid advancement of machine learned model development threatens this paradigm, as it can be difficult to understand how transferability is embedded (or missed) in complex models. While transferability in general chemistry machine learning should benefit from diverse training data, a rigorous understanding of transferability together with its interplay with chemical representation remains an open problem. We introduce a transferability framework and apply it to a controllable data-driven model for developing density functional approximations (DFAs), an indispensable tool in everyday chemistry research. We reveal that human intuition introduces chemical biases that can hamper the transferability of data-driven DFAs, and we identify strategies for their elimination. We then show that uncritical use of large training sets can actually hinder the transferability of DFAs, in contradiction to typical “more is more†expectations. Finally, our transferability framework yields transferable diversity, a cornerstone principle for data curation for developing general-purpose machine learning models in chemistry},
langid = {english},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AML,ML,todo-tagging,transfer learning},
file = {/Users/wasmer/Nextcloud/Zotero/Vuckovic et al_2023_Transferable diversity – a data-driven representation of chemical space.pdf}
}
@@ -16648,7 +18143,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
urldate = {2021-10-19},
abstract = {Accurate approximations to density functionals have recently been obtained via machine learning (ML). By applying ML to a simple function of one variable without any random sampling, we extract the qualitative dependence of errors on hyperparameters. We find universal features of the behavior in extreme limits, including both very small and very large length scales, and the noise-free limit. We show how such features arise in ML models of density functionals. © 2015 Wiley Periodicals, Inc.},
langid = {english},
- keywords = {\_tablet,density functional theory,DFT,extreme behaviors,hyperparameters optimization,KRR,machine learning,ML,models,noise-free curve,tutorial},
+ keywords = {density functional theory,DFT,extreme behaviors,hyperparameters optimization,KRR,machine learning,ML,models,noise-free curve,tutorial},
file = {/home/johannes/Nextcloud/Zotero/false;/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Vu et al_2015_Understanding kernel ridge regression.pdf;/Users/wasmer/Zotero/storage/5INUIEQC/qua.html}
}
@@ -16666,18 +18161,54 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
annotation = {OCLC: 633422775}
}
+@article{walterCorrectionSolarWater2011,
+ title = {Correction to {{Solar Water Splitting Cells}}},
+ author = {Walter, Michael G. and Warren, Emily L. and McKone, James R. and Boettcher, Shannon W. and Mi, Qixi and Santori, Elizabeth A. and Lewis, Nathan S.},
+ date = {2011-09-14},
+ journaltitle = {Chemical Reviews},
+ shortjournal = {Chem. Rev.},
+ volume = {111},
+ number = {9},
+ pages = {5815--5815},
+ publisher = {American Chemical Society},
+ issn = {0009-2665},
+ doi = {10.1021/cr200102n},
+ url = {https://doi.org/10.1021/cr200102n},
+ urldate = {2024-08-01},
+ keywords = {/unread,artificial photosynthesis,chemistry,energy challenge,energy materials,materials,solar energy,water splitting},
+ file = {/Users/wasmer/Nextcloud/Zotero/Walter et al_2011_Correction to Solar Water Splitting Cells.pdf}
+}
+
+@article{walterSolarWaterSplitting2010,
+ title = {Solar {{Water Splitting Cells}}},
+ author = {Walter, Michael G. and Warren, Emily L. and McKone, James R. and Boettcher, Shannon W. and Mi, Qixi and Santori, Elizabeth A. and Lewis, Nathan S.},
+ date = {2010-11-10},
+ journaltitle = {Chemical Reviews},
+ shortjournal = {Chem. Rev.},
+ volume = {110},
+ number = {11},
+ pages = {6446--6473},
+ publisher = {American Chemical Society},
+ issn = {0009-2665},
+ doi = {10.1021/cr1002326},
+ url = {https://doi.org/10.1021/cr1002326},
+ urldate = {2024-08-01},
+ keywords = {/unread,artificial photosynthesis,energy challenge,energy materials,materials,water splitting},
+ file = {/Users/wasmer/Nextcloud/Zotero/Walter et al_2010_Solar Water Splitting Cells.pdf}
+}
+
@online{wangApproximatelyEquivariantNetworks2022,
title = {Approximately {{Equivariant Networks}} for {{Imperfectly Symmetric Dynamics}}},
author = {Wang, Rui and Walters, Robin and Yu, Rose},
date = {2022-06-16},
eprint = {2201.11969},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2201.11969},
url = {http://arxiv.org/abs/2201.11969},
urldate = {2023-11-11},
abstract = {Incorporating symmetry as an inductive bias into neural network architecture has led to improvements in generalization, data efficiency, and physical consistency in dynamics modeling. Methods such as CNNs or equivariant neural networks use weight tying to enforce symmetries such as shift invariance or rotational equivariance. However, despite the fact that physical laws obey many symmetries, real-world dynamical data rarely conforms to strict mathematical symmetry either due to noisy or incomplete data or to symmetry breaking features in the underlying dynamical system. We explore approximately equivariant networks which are biased towards preserving symmetry but are not strictly constrained to do so. By relaxing equivariance constraints, we find that our models can outperform both baselines with no symmetry bias and baselines with overly strict symmetry in both simulated turbulence domains and real-world multi-stream jet flow.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {alternative approaches,approximat,approximative equivariance,CNN,data equivariance,Deep learning,equivariant,General ML,GNN,inductive bias,invariance,ML,model equivariance,symmetry},
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2022_Approximately Equivariant Networks for Imperfectly Symmetric Dynamics.pdf;/Users/wasmer/Zotero/storage/L7W3IW67/2201.html}
}
@@ -16688,14 +18219,14 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {Wang, Yuxiang and Li, He and Tang, Zechen and Tao, Honggeng and Wang, Yanzhen and Yuan, Zilong and Chen, Zezhou and Duan, Wenhui and Xu, Yong},
date = {2024-01-30},
eprint = {2401.17015},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2401.17015},
url = {http://arxiv.org/abs/2401.17015},
urldate = {2024-04-18},
abstract = {Deep-learning electronic structure calculations show great potential for revolutionizing the landscape of computational materials research. However, current neural-network architectures are not deemed suitable for widespread general-purpose application. Here we introduce a framework of equivariant local-coordinate transformer, designed to enhance the deep-learning density functional theory Hamiltonian referred to as DeepH-2. Unlike previous models such as DeepH and DeepH-E3, DeepH-2 seamlessly integrates the simplicity of local-coordinate transformations and the mathematical elegance of equivariant neural networks, effectively overcoming their respective disadvantages. Based on our comprehensive experiments, DeepH-2 demonstrates superiority over its predecessors in both efficiency and accuracy, showcasing state-of-the-art performance. This advancement opens up opportunities for exploring universal neural network models or even large materials models.},
- pubstate = {preprint},
- keywords = {\_tablet,AML,DeepH,ELCT,Equiformer,equivariant,LCNN,local coordinates,ML,ML-DFT,ML-ESM,prediction of Hamiltonian matrix,transformer,with-code},
+ pubstate = {prepublished},
+ keywords = {AML,DeepH,ELCT,Equiformer,equivariant,LCNN,local coordinates,ML,ML-DFT,ML-ESM,prediction of Hamiltonian matrix,transformer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2024_DeepH-2.pdf;/Users/wasmer/Zotero/storage/WVSPJ8YJ/2401.html}
}
@@ -16704,13 +18235,13 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {Wang, Yuyang and Elhag, Ahmed A. and Jaitly, Navdeep and Susskind, Joshua M. and Bautista, Miguel Angel},
date = {2023-11-27},
eprint = {2311.17932},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2311.17932},
url = {http://arxiv.org/abs/2311.17932},
urldate = {2023-12-04},
abstract = {In this paper we tackle the problem of generating conformers of a molecule in 3D space given its molecular graph. We parameterize these conformers as continuous functions that map elements from the molecular graph to points in 3D space. We then formulate the problem of learning to generate conformers as learning a distribution over these functions using a diffusion generative model, called Molecular Conformer Fields (MCF). Our approach is simple and scalable, and achieves state-of-the-art performance on challenging molecular conformer generation benchmarks while making no assumptions about the explicit structure of molecules (e.g. modeling torsional angles). MCF represents an advance in extending diffusion models to handle complex scientific problems in a conceptually simple, scalable and effective manner.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread},
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2023_Generating Molecular Conformer Fields.pdf;/Users/wasmer/Zotero/storage/HPM6A9N3/2311.html}
}
@@ -16720,13 +18251,13 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {Wang, Yuanqing and Fass, Josh and Stern, Chaya D. and Luo, Kun and Chodera, John},
date = {2019-09-17},
eprint = {1909.07903},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.1909.07903},
url = {http://arxiv.org/abs/1909.07903},
urldate = {2022-09-27},
abstract = {Atomic partial charges are crucial parameters for Molecular Dynamics (MD) simulations, molecular mechanics calculations, and virtual screening, as they determine the electrostatic contributions to interaction energies. Current methods for calculating partial charges, however, are either slow and scale poorly with molecular size (quantum chemical methods) or unreliable (empirical methods). Here, we present a new charge derivation method based on Graph Nets---a set of update and aggregate functions that operate on molecular topologies and propagate information thereon---that could approximate charges derived from Density Functional Theory (DFT) calculations with high accuracy and an over 500-fold speed up.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {GCN,GNN,molecules,prediction of partial charge},
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2019_Graph Nets for Partial Charge Prediction.pdf;/Users/wasmer/Zotero/storage/5MD2WVP3/1909.html}
}
@@ -16756,13 +18287,13 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {Wang, Yuan and Zhang, Fayuan and Zeng, Meng and Sun, Hongyi and Hao, Zhanyang and Cai, Yongqing and Rong, Hongtao and Zhang, Chengcheng and Liu, Cai and Ma, Xiaoming and Wang, Le and Guo, Shu and Lin, Junhao and Liu, Qihang and Liu, Chang and Chen, Chaoyu},
date = {2022-12-18},
eprint = {2212.09057},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2212.09057},
url = {http://arxiv.org/abs/2212.09057},
urldate = {2023-06-15},
abstract = {Topological states of matter possess bulk electronic structures categorized by topological invariants and edge/surface states due to the bulk-boundary correspondence. Topological materials hold great potential in the development of dissipationless spintronics, information storage, and quantum computation, particularly if combined with magnetic order intrinsically or extrinsically. Here, we review the recent progress in the exploration of intrinsic magnetic topological materials, including but not limited to magnetic topological insulators, magnetic topological metals, and magnetic Weyl semimetals. We pay special attention to their characteristic band features such as the gap of topological surface state, gapped Dirac cone induced by magnetization (either bulk or surface), Weyl nodal point/line, and Fermi arc, as well as the exotic transport responses resulting from such band features. We conclude with a brief envision for experimental explorations of new physics or effects by incorporating other orders in intrinsic magnetic topological materials.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,magnetic materials,magnetic topological materials,topological},
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2022_Intrinsic Magnetic Topological Materials.pdf;/Users/wasmer/Zotero/storage/MLELX7M9/2212.html}
}
@@ -16807,6 +18338,20 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2022_Large scale dataset of real space electronic charge density of cubic inorganic.pdf}
}
+@article{wangMachineLearningAccelerated2023,
+ title = {Machine {{Learning Accelerated Study}} of {{Defect Energy Levels}} in {{Perovskites}}},
+ author = {Wang, Jianwei and Chen, Haiyuan and Wu, Xiaoyu and Niu, Xiaobin},
+ date = {2023-06-06},
+ journaltitle = {The Journal of Physical Chemistry C},
+ volume = {127},
+ doi = {10.1021/acs.jpcc.3c02493},
+ url = {https://discovery.researcher.life/article/machine-learning-accelerated-study-of-defect-energy-levels-in-perovskites/d3ff764ec06c3f74bb47e79f0c08e721},
+ urldate = {2024-05-25},
+ abstract = {Manipulating the defect tolerance is one of the effective ways to maintain the high power conversion efficiency and keep the stability of perovskite semiconductor materials. So, rapid screening for defects and trap states in the perovskite semiconductor candidates is urgently needed. Theoretical investigations of defects based on density functional theory (DFT) are still limited by their extremely high consumption of computational resources and time. We implement an accelerated material discovery approach using artificial intelligence and DFT, which can predict the defect transition levels in the candidate perovskite semiconductor materials. To verify the accuracy of our models, Cs3Sb2Br9 and Cs2SnBr6, which are out of the dataset that we used in machine learning (ML) model construction, are taken as examples. The extrapolation of ML prediction models and the results given by DFT calculations are compared for defect transition energy levels. The two methods are consistent with each other with very small errors. Our strategy avoids complex and time-consuming computational work based on DFT and provides quick and efficient screening of physical properties with low cost.},
+ langid = {english},
+ keywords = {/unread}
+}
+
@article{wangMachineLearningMaterials2020,
title = {Machine {{Learning}} for {{Materials Scientists}}: {{An Introductory Guide}} toward {{Best Practices}}},
shorttitle = {Machine {{Learning}} for {{Materials Scientists}}},
@@ -16888,7 +18433,7 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
abstract = {Artificial intelligence (AI) is being increasingly integrated into scientific discovery to augment and accelerate research, helping scientists to generate hypotheses, design experiments, collect and interpret large datasets, and gain insights that might not have been possible using traditional scientific methods alone. Here we examine breakthroughs over the past decade that include self-supervised learning, which allows models to be trained on vast amounts of unlabelled data, and geometric deep learning, which leverages knowledge about the structure of scientific data to enhance model accuracy and efficiency. Generative AI methods can create designs, such as small-molecule drugs and proteins, by analysing diverse data modalities, including images and sequences. We discuss how these methods can help scientists throughout the scientific process and the central issues that remain despite such advances. Both developers and users of AI tools need a better understanding of when such approaches need improvement, and challenges posed by poor data quality and stewardship remain. These issues cut across scientific disciplines and require developing foundational algorithmic approaches that can contribute to scientific understanding or acquire it autonomously, making them critical areas of focus for AI innovation.},
issue = {7972},
langid = {english},
- keywords = {AI4Science,equivariant,General ML,geometric deep learning,LLM,ML,MPNN,neural operator,review,review-of-AI4science,roadmap,symmetry,transformer},
+ keywords = {AI4Science,biomolecules,chemical reaction,chemistry,database generation,equivariant,General ML,geometric deep learning,LLM,ML,MPNN,neural operator,review,review-of-AI4science,roadmap,symmetry,transformer,weather forecasting},
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2023_Scientific discovery in the age of artificial intelligence.pdf}
}
@@ -16915,14 +18460,14 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
author = {Wang, Hai-Chen and Schmidt, Jonathan and Marques, Miguel A. L. and Wirtz, Ludger and Romero, Aldo H.},
date = {2022-12-07},
eprint = {2212.03975},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2212.03975},
url = {http://arxiv.org/abs/2212.03975},
urldate = {2023-04-04},
abstract = {We present a symmetry-based exhaustive approach to explore the structural and compositional richness of two-dimensional materials. We use a ``combinatorial engine'' that constructs potential compounds by occupying all possible Wyckoff positions for a certain space group with combinations of chemical elements. These combinations are restricted by imposing charge neutrality and the Pauling test for electronegativities. The structures are then pre-optimized with a specially crafted universal neural-network force-field, before a final step of geometry optimization using density-functional theory is performed. In this way we unveil an unprecedented variety of two-dimensional materials, covering the whole periodic table in more than 30 different stoichiometries of form A\$\_n\$B\$\_m\$ or A\$\_n\$B\$\_m\$C\$\_k\$. Among the found structures we find examples that can be built by decorating nearly all Platonic and Archimedean tesselations as well as their dual Laves or Catalan tilings. We also obtain a rich, and unexpected, polymorphism for some specific compounds. We further accelerate the exploration of the chemical space of two-dimensional materials by employing machine-learning-accelerated prototype search, based on the structural types discovered in the exhaustive search. In total, we obtain around 6500 compounds, not present in previous available databases of 2D materials, with an energy of less than 250\textasciitilde meV/atom above the convex hull of thermodynamic stability.},
- pubstate = {preprint},
- keywords = {\_tablet,2D material,2DMatpedia,AML,binary systems,C2DB,CGAT,convex hull,crystal structure,crystal symmetry,M3GNet,materials discovery,materials screening,MC2D,ML,polymorphs,prediction of structure,ternary systems,thermodynamic stability,V2DB,Wyckoff positions},
+ pubstate = {prepublished},
+ keywords = {2D material,2DMatpedia,AML,binary systems,C2DB,CGAT,convex hull,crystal structure,crystal symmetry,M3GNet,materials discovery,materials screening,MC2D,ML,polymorphs,prediction of structure,ternary systems,thermodynamic stability,V2DB,Wyckoff positions},
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2022_Symmetry-based computational search for novel binary and ternary 2D materials.pdf;/Users/wasmer/Zotero/storage/I7GDFM7H/2212.html}
}
@@ -16941,7 +18486,6 @@ Subject\_term\_id: electronic-devices;electronic-properties-and-materials;ferroe
urldate = {2023-06-14},
abstract = {Topological quantum materials are a class of compounds featuring electronic band structures, which are topologically distinct from common metals and insulators. These materials have emerged as exceptionally fertile ground for materials science research. The topologically nontrivial electronic structures of these materials support many interesting properties, ranging from the topologically protected states, manifesting as high mobility and spin-momentum locking, to various quantum Hall effects, axionic physics, and Majorana modes. In this article, we describe different topological matters, including topological insulators, Weyl semimetals, twisted graphene, and related two-dimensional Chern magnetic insulators, as well as their heterostructures. We focus on recent materials discoveries and experimental advancements of topological materials, and their heterostructures. Finally, we conclude with prospects for the discovery of additional topological materials for studying quantum processes, quasiparticles and their composites, as well as exploiting potential applications of these materials.},
langid = {english},
- keywords = {\_tablet},
file = {/Users/wasmer/Nextcloud/Zotero/Wang et al_2020_Topological quantum materials.pdf}
}
@@ -17046,15 +18590,30 @@ Subject\_term\_id: electronic-properties-and-materials;quantum-hall;superconduct
file = {/Users/wasmer/Nextcloud/Zotero/Waroquiers et al_2017_Statistical Analysis of Coordination Environments in Oxides.pdf}
}
-@software{wasmerAiidajutools2022,
- title = {Aiida-Jutools},
- author = {Wasmer, Johannes},
- date = {2022-01-19T08:36:40Z},
- origdate = {2020-01-08T15:38:01Z},
- url = {https://github.com/JuDFTteam/aiida-jutools},
- urldate = {2022-10-16},
- abstract = {Tools for managing high-throughput experiments with AiiDA.},
- organization = {JuDFTteam}
+@online{wasmerBestAtomisticMachine2023,
+ title = {Best of {{Atomistic Machine Learning}}},
+ author = {Wasmer, Johannes and Evans, Matthew and Blaiszik, Ben and Riebesell, Janosh},
+ date = {2023-12-25},
+ doi = {10.5281/zenodo.10934602},
+ url = {https://doi.org/10.5281/zenodo.10934602},
+ urldate = {2024-05-29},
+ abstract = {A ranked list of awesome atomistic machine learning projects.},
+ keywords = {/unread,ai4science,atomistic-machine-learning,awesome-list,best-of-list,chemistry-datasets,community-resource,computational-chemistry,computational-materials-science,density-functional-theory,drug-discovery,electronic-structure,interatomic-potentials,living-document,materials-datasets,materials-discovery,materials-informatics,molecular-dynamics,quantum-chemistry,scientific-machine-learning,surrogate-models}
+}
+
+@report{wasmerBestAtomisticMachine2023a,
+ title = {Best of {{Atomistic Machine Learning}}},
+ author = {Wasmer, Johannes and Riebesell, Janosh and Evans, Matthew and Blaiszik, Ben},
+ date = {2023},
+ number = {FZJ-2023-05862},
+ institution = {Quanten-Theorie der Materialien},
+ doi = {10.5281/ZENODO.10430261},
+ url = {https://juser.fz-juelich.de/record/1020061},
+ urldate = {2024-06-05},
+ abstract = {A ranked list of awesome atomistic machine learning projects. Wasmer, Johannes; Evans, Matthew; Blaiszik, Ben; Riebesell, Janosh},
+ langid = {english},
+ keywords = {/unread,AML,FZJ,ML,PGI,PGI-1/IAS-1,review,with-code},
+ file = {/Users/wasmer/Zotero/storage/7N6RJXIT/1020061.html}
}
@unpublished{wasmerComparisonStructuralRepresentations2021,
@@ -17081,23 +18640,37 @@ Subject\_term\_id: electronic-properties-and-materials;quantum-hall;superconduct
urldate = {2022-08-08},
langid = {english},
pagetotal = {99},
- keywords = {\_tablet,AiiDA,aiida-kkr,AML,Coulomb matrix,descriptor comparison,impurity embedding,juKKR,KKR,master-thesis,ML,PGI-1/IAS-1,SOAP,thesis},
+ keywords = {AiiDA,aiida-kkr,AML,Coulomb matrix,descriptor comparison,impurity embedding,juKKR,KKR,master-thesis,ML,PGI-1/IAS-1,SOAP,thesis},
file = {/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Wasmer_2021_Development of a surrogate machine learning model for the acceleration of.pdf;/Users/wasmer/Zotero/storage/AC483X2N/master-thesis.html}
}
+@thesis{wasmerDevelopmentSurrogateMachine2021a,
+ title = {Development of a Surrogate Machine Learning Model for the Acceleration of Density Functional Calculations with the {{Korringa-Kohn-Rostoker}} Method},
+ author = {Wasmer, Johannes and Berkels, Benjamin and Rüssmann, Philipp and Blügel, Stefan},
+ date = {2021},
+ number = {FZJ-2023-05854},
+ institution = {Masterarbeit, RWTH Aachen University, 2022},
+ url = {https://juser.fz-juelich.de/record/1020053},
+ urldate = {2024-06-05},
+ abstract = {Density functional theory (DFT) has become an indispensable tool in materials science. Specialized DFT methods like the Korringa-Kohan Rostoker Green Function (KKR) method are predestined to investigate the technologically relevant effects of crystallographic defects on the electronic and magnetic structure of host materials. This thesis lays the groundwork for answering the question of whether surrogate machine learning (ML) models have the potential to accelerate such DFT calculations since their computational complexity severely limits them to systems sizes of about a thousand atoms in practice. To that end, a versatile suite of software tools that facilitates the generation and analysis of high-throughput computing DFT datasets with the JuKKR DFT codes and the AiiDA workflow engine is presented. We demonstrate its use by generating a database of 8,760 converged KKR DFT calculations of single impurity embeddings into elemental crystals with 60 different chemical elements and varying lattice constants and that preserves the full data provenance of each calculation. Finally, we use the single-impurity database to compare the Coulomb Matrix and the Smooth Overlap of Atomic Positions (SOAP) as structural descriptors of the local atomic environment for materials defects. Their potential use in surrogate ML models is showcased in a simple example of host crystal structure prediction that achieves 93 percent accuracy. Wasmer, Johannes; Blügel, Stefan; Rüssmann, Philipp; Berkels, Benjamin},
+ langid = {english},
+ keywords = {/unread},
+ file = {/Users/wasmer/Zotero/storage/NTW3FHQ6/1020053.html}
+}
+
@online{weiGraphLearningIts2023,
title = {Graph {{Learning}} and {{Its Applications}}: {{A Holistic Survey}}},
shorttitle = {Graph {{Learning}} and {{Its Applications}}},
author = {Wei, Shaopeng and Zhao, Yu and Chen, Xingyan and Li, Qing and Zhuang, Fuzhen and Liu, Ji and Kou, Gang},
date = {2023-06-03},
eprint = {2212.08966},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2212.08966},
url = {http://arxiv.org/abs/2212.08966},
urldate = {2023-11-14},
abstract = {Graph learning is a prevalent domain that endeavors to learn the intricate relationships among nodes and the topological structure of graphs. Over the years, graph learning has transcended from graph theory to graph data mining. With the advent of representation learning, it has attained remarkable performance in diverse scenarios. Owing to its extensive application prospects, graph learning attracts copious attention. While some researchers have accomplished impressive surveys on graph learning, they failed to connect related objectives, methods, and applications in a more coherent way. As a result, they did not encompass current ample scenarios and challenging problems due to the rapid expansion of graph learning. Particularly, large language models have recently had a disruptive effect on human life, but they also show relative weakness in structured scenarios. The question of how to make these models more powerful with graph learning remains open. Different from previous surveys on graph learning, we provide a holistic review that analyzes current works from the perspective of graph structure, and discusses the latest applications, trends, and challenges in graph learning. Specifically, we commence by proposing a taxonomy and then summarize the methods employed in graph learning. We then provide a detailed elucidation of mainstream applications. Finally, we propose future directions.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {Computer Science - Artificial Intelligence},
file = {/Users/wasmer/Nextcloud/Zotero/Wei et al_2023_Graph Learning and Its Applications.pdf;/Users/wasmer/Zotero/storage/WLJFND3J/2212.html}
}
@@ -17111,7 +18684,7 @@ Subject\_term\_id: electronic-properties-and-materials;quantum-hall;superconduct
urldate = {2023-11-18},
abstract = {In this book, Equivariant and Coordinate Independent Convolutional Networks, we develop a gauge theory of artificial neural networks for processing spatially structured data like images, audio, or videos. The standard neural network architecture for such data are convolutional networks, which are characterized by their position-independent inference. Generalizing whatever they learn over spatial locations, convolutional networks are substantially more data efficient and robust in comparison to non-convolutional models. This characteristic is especially important in domains like medical imaging, where training data is scarce. The independence from spatial locations is formally captured by the networks’ translation group equivariance, i.e. their property to commute with translations of their input signals. We show that the convolutional network design is not only sufficient for translation equivariance but is actually a necessary condition – convolutions can therefore be derived by demanding the model’s equivariance. The first part of this work leverages this insight to define generalized convolutional networks which are equivariant under larger symmetry groups. Such models generalize their inference over additional geometric transformations, for instance, rotations or reflections of patterns in images. We demonstrate empirically that they exhibit a significantly enhanced data efficiency, convergence rate, and final performance in comparison to conventional convolutional networks. Our publicly available implementation found wide use in the research community. In the second part, we extend convolutional networks further to process signals on Riemannian manifolds. Beyond flat Euclidean images, this setting includes, e.g., spherical signals like global weather patterns on the earth’s surface, or signals on general surfaces like artery walls or the cerebral cortex. We show that convolution kernels on manifolds are required to be equivariant under local gauge transformations if the networks’ inference is demanded to be coordinate independent. The resulting coordinate independent networks are proven to be equivariant with respect to the manifolds’ global symmetries (isometries). Our objective is not to propose yet another equivariant network design for a narrow application domain, but to devise a unifying mathematical framework for convolutional networks. The last part of this book demonstrates the generality of our differential geometric formulation of convolutional networks by showing that is able to explain a vast number of equivariant network architectures from the literature.},
langid = {american},
- keywords = {\_tablet,CNN,covariant,educational,equivariant,gauge theory,General ML,geometric deep learning,GNN,group theory,invariance,learning material,ML,ML theory,online book,physics-informed ML,review,review-of-GDL,steerable CNN,symmetry,textbook},
+ keywords = {CNN,covariant,educational,equivariant,gauge theory,General ML,geometric deep learning,GNN,group theory,invariance,learning material,ML,ML theory,online book,physics-informed ML,review,review-of-GDL,steerable CNN,symmetry,textbook},
file = {/Users/wasmer/Nextcloud/Zotero/Weiler et al_2023_Equivariant and Coordinate Independent Convolutional Networks - A Gauge Field.pdf;/Users/wasmer/Zotero/storage/U2AEW2RU/maurice-weiler.gitlab.io.html}
}
@@ -17120,13 +18693,13 @@ Subject\_term\_id: electronic-properties-and-materials;quantum-hall;superconduct
author = {Weiler, Maurice and Cesa, Gabriele},
date = {2021-04-06},
eprint = {1911.08251},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, eess},
doi = {10.48550/arXiv.1911.08251},
url = {http://arxiv.org/abs/1911.08251},
urldate = {2023-06-30},
abstract = {The big empirical success of group equivariant networks has led in recent years to the sprouting of a great variety of equivariant network architectures. A particular focus has thereby been on rotation and reflection equivariant CNNs for planar images. Here we give a general description of \$E(2)\$-equivariant convolutions in the framework of Steerable CNNs. The theory of Steerable CNNs thereby yields constraints on the convolution kernels which depend on group representations describing the transformation laws of feature spaces. We show that these constraints for arbitrary group representations can be reduced to constraints under irreducible representations. A general solution of the kernel space constraint is given for arbitrary representations of the Euclidean group \$E(2)\$ and its subgroups. We implement a wide range of previously proposed and entirely new equivariant network architectures and extensively compare their performances. \$E(2)\$-steerable convolutions are further shown to yield remarkable gains on CIFAR-10, CIFAR-100 and STL-10 when used as a drop-in replacement for non-equivariant convolutions.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,2D,CNN,computer vision,E(2),equivariant,General ML,library,ML,steerable CNN,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Weiler_Cesa_2021_General $E(2)$-Equivariant Steerable CNNs.pdf;/Users/wasmer/Zotero/storage/49VD5PUN/1911.html}
}
@@ -17178,8 +18751,8 @@ Subject\_term\_id: electronic-properties-and-materials;quantum-hall;superconduct
urldate = {2022-12-29},
abstract = {Chemists can be skeptical in using deep learning (DL) in decision making, due to the lack of interpretability in "black-box" models. Explainable artificial intelligence (XAI) is a branch of AI which addresses this drawback by providing tools to interpret DL models and their predictions. We review the principles of XAI in the domain of chemistry and emerging methods for creating and evaluating explanations. Then we focus methods developed by our group and their application to predicting solubility, blood-brain barrier permeability, and the scent of molecules. We show that XAI methods like chemical counterfactuals and descriptor explanations can both explain DL predictions and give insight into structure-property relationships. Finally, we discuss how a two step process of highly accurate black-box modeling and then creating explanations gives both highly accurate predictions and clear structure-property relationships.},
langid = {english},
- pubstate = {preprint},
- keywords = {\_tablet,counterfactual explanation,Deep learning,GNN,molecules,XAI},
+ pubstate = {prepublished},
+ keywords = {counterfactual explanation,Deep learning,GNN,molecules,XAI},
file = {/Users/wasmer/Nextcloud/Zotero/Wellawatte et al_2022_A Perspective on Explanations of Molecular Prediction Models.pdf}
}
@@ -17239,7 +18812,7 @@ Subject\_term\_id: electronic-properties-and-materials;quantum-hall;superconduct
author = {Westermayr, Julia and Gastegger, Michael and Schütt, Kristof T. and Maurer, Reinhard J.},
date = {2021-04-20},
eprint = {2102.08435},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
url = {http://arxiv.org/abs/2102.08435},
urldate = {2021-05-13},
@@ -17283,7 +18856,7 @@ Subject\_term\_id: electronic-properties-and-materials;quantum-hall;superconduct
abstract = {Deep learning is becoming a standard tool in chemistry and materials science. Although there are learning materials available for deep learning, none cover the applications in chemistry and materials science or the peculiarities of working with molecules. The textbook described here provides a systematic and applied introduction to the latest research in deep learning in chemistry and materials science. It covers the math fundamentals, the requisite machine learning, the common neural network architectures used today, and the details necessary to be a practitioner of deep learning. The textbook is a living document and will be updated as the rapidly changing deep learning field evolves.},
issue = {1},
langid = {english},
- keywords = {\_tablet,book,GNN,ML-DFT,ML-ESM,MLP,MPNN,tutorial},
+ keywords = {book,GNN,ML-DFT,ML-ESM,MLP,MPNN,tutorial},
file = {/Users/wasmer/Nextcloud/Zotero/White_2021_Deep Learning for Molecules and Materials.pdf}
}
@@ -17331,10 +18904,18 @@ Subject\_term\_id: electronic-properties-and-materials;quantum-hall;superconduct
urldate = {2022-09-27},
abstract = {As graph neural networks are becoming more and more powerful and useful in the field of drug discovery, many pharmaceutical companies are getting interested in utilizing these methods for their own in-house frameworks. This is especially compelling for tasks such as the prediction of molecular properties which is often one of the most crucial tasks in computer-aided drug discovery workflows. The immense hype surrounding these kinds of algorithms has led to the development of many different types of promising architectures and in this review we try to structure this highly dynamic field of AI-research by collecting and classifying 80 GNNs that have been used to predict more than 20 molecular properties using 48 different datasets.},
langid = {english},
- keywords = {\_tablet,chemistry,GCN,GNN,molecules,review},
+ keywords = {chemistry,GCN,GNN,molecules,review},
file = {/Users/wasmer/Nextcloud/Zotero/Wieder et al_2020_A compact review of molecular property prediction with graph neural networks.pdf;/Users/wasmer/Zotero/storage/KHCYV2ZB/S1740674920300305.html}
}
+@online{WilhelmUndElse,
+ title = {Wilhelm Und {{Else Heraeus-Stiftung}}: 787. {{WE-Heraeus-Seminar}}},
+ url = {https://www.we-heraeus-stiftung.de/veranstaltungen/accelerated-discovery-of-new-materials/},
+ urldate = {2024-08-01},
+ keywords = {/unread,AML,conference,energy challenge,energy materials,materials acceleration platforms},
+ file = {/Users/wasmer/Zotero/storage/YVCVEXE9/accelerated-discovery-of-new-materials.html}
+}
+
@article{wilkinsAccurateMolecularPolarizabilities2019,
title = {Accurate Molecular Polarizabilities with Coupled Cluster Theory and Machine Learning},
author = {Wilkins, David M. and Grisafi, Andrea and Yang, Yang and Lao, Ka Un and DiStasio, Robert A. and Ceriotti, Michele},
@@ -17421,7 +19002,7 @@ Subject\_term\_id: publication-characteristics;research-data},
author = {Winter, Robin and Bertolini, Marco and Le, Tuan and Noé, Frank and Clevert, Djork-Arné},
date = {2022-02-15},
eprint = {2202.07559},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
url = {http://arxiv.org/abs/2202.07559},
urldate = {2022-05-11},
@@ -17436,13 +19017,13 @@ Subject\_term\_id: publication-characteristics;research-data},
author = {Witt, William C. and family=Oord, given=Cas, prefix=van der, useprefix=true and Gelžinytė, Elena and Järvinen, Teemu and Ross, Andres and Darby, James P. and Ho, Cheuk Hin and Baldwin, William J. and Sachs, Matthias and Kermode, James and Bernstein, Noam and Csányi, Gábor and Ortner, Christoph},
date = {2023-09-07},
eprint = {2309.03161},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2309.03161},
url = {http://arxiv.org/abs/2309.03161},
urldate = {2023-09-22},
abstract = {We introduce ACEpotentials.jl, a Julia-language software package that constructs interatomic potentials from quantum mechanical reference data using the Atomic Cluster Expansion (Drautz, 2019). As the latter provides a complete description of atomic environments, including invariance to overall translation and rotation as well as permutation of like atoms, the resulting potentials are systematically improvable and data efficient. Furthermore, the descriptor's expressiveness enables use of a linear model, facilitating rapid evaluation and straightforward application of Bayesian techniques for active learning. We summarize the capabilities of ACEpotentials.jl and demonstrate its strengths (simplicity, interpretability, robustness, performance) on a selection of prototypical atomistic modelling workflows.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ACE,active learning,AML,descriptors,invariance,Julia,library,linear regression,ML,sparsification,todo-tagging,tutorial,with-code},
file = {/Users/wasmer/Zotero/storage/P4J69KC5/Witt et al. - 2023 - ACEpotentials.jl A Julia Implementation of the At.pdf;/Users/wasmer/Zotero/storage/SCPAUKAV/2309.html}
}
@@ -17472,14 +19053,14 @@ Subject\_term\_id: publication-characteristics;research-data},
author = {Woodgate, Christopher D. and Hedlund, Daniel and Lewis, L. H. and Staunton, Julie B.},
date = {2023-03-01},
eprint = {2303.00641},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2303.00641},
url = {http://arxiv.org/abs/2303.00641},
urldate = {2023-04-04},
abstract = {The impact of magnetism on predicted atomic short-range order in Ni-based high-entropy alloys is studied using a first-principles, all-electron, Landau-type linear response theory, coupled with lattice-based atomistic modelling. We perform two sets of linear-response calculations: one in which the paramagnetic state is modelled within the disordered local moment picture, and one in which systems are modelled in a magnetically ordered state. We show that the treatment of magnetism can have significant impact both on the predicted temperature of atomic ordering and also the nature of atomic order itself. In CrCoNi, we find that the nature of atomic order changes from being \$L1\_2\$-like when modelled in the paramagnetic state to MoPt\$\_2\$-like when modelled assuming the system has magnetically ordered. In CrFeCoNi, atomic correlations between Fe and the other elements present are dramatically strengthened when we switch from treating the system as magnetically disordered to magnetically ordered. Our results show it is necessary to consider the magnetic state when modelling multicomponent alloys containing mid- to late-\$3d\$ elements. Further, we suggest that there may be high-entropy alloy compositions containing \$3d\$ transition metals that will exhibit specific atomic short-range order when thermally treated in an applied magnetic field.},
- pubstate = {preprint},
- keywords = {/unread,\_tablet,CPA,DFT,high-entropy alloys,KKR,n-ary alloys,transition metals},
+ pubstate = {prepublished},
+ keywords = {/unread,CPA,DFT,high-entropy alloys,KKR,n-ary alloys,transition metals},
file = {/Users/wasmer/Nextcloud/Zotero/Woodgate et al_2023_Interplay between magnetism and short-range order in Ni-based high-entropy.pdf;/Users/wasmer/Zotero/storage/RAQ822ZS/2303.html}
}
@@ -17516,13 +19097,13 @@ Subject\_term\_id: publication-characteristics;research-data},
author = {Wu, Zhenqin and Ramsundar, Bharath and Feinberg, Evan N. and Gomes, Joseph and Geniesse, Caleb and Pappu, Aneesh S. and Leswing, Karl and Pande, Vijay},
date = {2018-10-25},
eprint = {1703.00564},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics, stat},
doi = {10.48550/arXiv.1703.00564},
url = {http://arxiv.org/abs/1703.00564},
urldate = {2023-09-25},
abstract = {Molecular machine learning has been maturing rapidly over the last few years. Improved methods and the presence of larger datasets have enabled machine learning algorithms to make increasingly accurate predictions about molecular properties. However, algorithmic progress has been limited due to the lack of a standard benchmark to compare the efficacy of proposed methods; most new algorithms are benchmarked on different datasets making it challenging to gauge the quality of proposed methods. This work introduces MoleculeNet, a large scale benchmark for molecular machine learning. MoleculeNet curates multiple public datasets, establishes metrics for evaluation, and offers high quality open-source implementations of multiple previously proposed molecular featurization and learning algorithms (released as part of the DeepChem open source library). MoleculeNet benchmarks demonstrate that learnable representations are powerful tools for molecular machine learning and broadly offer the best performance. However, this result comes with caveats. Learnable representations still struggle to deal with complex tasks under data scarcity and highly imbalanced classification. For quantum mechanical and biophysical datasets, the use of physics-aware featurizations can be more important than choice of particular learning algorithm.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AML,benchmark dataset,benchmarking,Database,graph ML,ML,molecules},
file = {/Users/wasmer/Zotero/storage/HQS5MMUL/Wu et al. - 2018 - MoleculeNet A Benchmark for Molecular Machine Lea.pdf;/Users/wasmer/Zotero/storage/RAREFZ8Y/1703.html}
}
@@ -17608,7 +19189,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
url = {https://link.aps.org/doi/10.1103/PhysRevLett.120.145301},
urldate = {2022-09-27},
abstract = {The use of machine learning methods for accelerating the design of crystalline materials usually requires manually constructed feature vectors or complex transformation of atom coordinates to input the crystal structure, which either constrains the model to certain crystal types or makes it difficult to provide chemical insights. Here, we develop a crystal graph convolutional neural networks framework to directly learn material properties from the connection of atoms in the crystal, providing a universal and interpretable representation of crystalline materials. Our method provides a highly accurate prediction of density functional theory calculated properties for eight different properties of crystals with various structure types and compositions after being trained with 104 data points. Further, our framework is interpretable because one can extract the contributions from local chemical environments to global properties. Using an example of perovskites, we show how this information can be utilized to discover empirical rules for materials design.},
- keywords = {\_tablet,CGCNN,GCN,GNN,library,solids,with-code},
+ keywords = {AML,CGCNN,crystal graph,GCN,GNN,library,materials,ML,MLP,original publication,prediction of energy,solids,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Xie_Grossman_2018_Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable.pdf}
}
@@ -17639,7 +19220,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
volume = {34},
number = {18},
eprint = {2104.11150},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
pages = {183002},
issn = {0953-8984, 1361-648X},
@@ -17647,7 +19228,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
url = {http://arxiv.org/abs/2104.11150},
urldate = {2023-05-06},
abstract = {Even though superconductivity has been studied intensively for more than a century, the vast majority of superconductivity research today is carried out in nearly the same manner as decades ago. That is, each study tends to focus on only a single material or small subset of materials, and discoveries are made more or less serendipitously. Recent increases in computing power, novel machine learning algorithms, and improved experimental capabilities offer new opportunities to revolutionize superconductor discovery. These will enable the rapid prediction of structures and properties of novel materials in an automated, high-throughput fashion and the efficient experimental testing of these predictions. Here, we review efforts to use machine learning to attain this goal.},
- keywords = {/unread,Allen-Dynes equation,AML,autoencoder,Eliashberg theory,hydrides,materials discovery,materials screening,ML,superconductor,symbolic regression},
+ keywords = {Allen-Dynes equation,AML,autoencoder,Eliashberg theory,hydrides,materials discovery,materials screening,ML,superconductor,symbolic regression},
file = {/Users/wasmer/Nextcloud/Zotero/Xie et al_2022_Towards high-throughput superconductor discovery via machine learning.pdf;/Users/wasmer/Zotero/storage/IVC6AHH2/2104.html}
}
@@ -17677,7 +19258,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Xie, Stephen R. and Rupp, Matthias and Hennig, Richard G.},
date = {2021-10-01},
eprint = {2110.00624},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
url = {http://arxiv.org/abs/2110.00624},
urldate = {2022-05-09},
@@ -17691,13 +19272,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Xie, Stephen R. and Rupp, Matthias and Hennig, Richard G.},
date = {2021-10-01},
eprint = {2110.00624},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2110.00624},
url = {http://arxiv.org/abs/2110.00624},
urldate = {2023-05-06},
abstract = {All-atom dynamics simulations are an indispensable quantitative tool in physics, chemistry, and materials science, but large systems and long simulation times remain challenging due to the trade-off between computational efficiency and predictive accuracy. To address this challenge, we combine effective two- and three-body potentials in a cubic B-spline basis with regularized linear regression to obtain machine-learning potentials that are physically interpretable, sufficiently accurate for applications, as fast as the fastest traditional empirical potentials, and two to four orders of magnitude faster than state-of-the-art machine-learning potentials. For data from empirical potentials, we demonstrate exact retrieval of the potential. For data from density functional theory, the predicted energies, forces, and derived properties, including phonon spectra, elastic constants, and melting points, closely match those of the reference method. The introduced potentials might contribute towards accurate all-atom dynamics simulations of large atomistic systems over long time scales.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,AML,B-splines,benchmarking,body-order,descriptors,GAP,library,linear regression,ML,MLP,MLP comparison,original publication,qSNAP,SNAP,UF3,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/false;/Users/wasmer/Zotero/storage/B9DGEUPF/2110.html}
}
@@ -17812,6 +19393,25 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
file = {/Users/wasmer/Nextcloud/Zotero/Yamashita_Sakuma_2022_First-Principles Study for Finite Temperature Magnetrocrystaline Anisotropy of.pdf}
}
+@article{yamashitaMagnetizationExchangestiffnessConstants2024,
+ title = {Magnetization and Exchange-Stiffness Constants of {{Fe}}–{{Al}}–{{Si}} Alloys at Finite Temperatures: {{A}} First-Principles Study},
+ shorttitle = {Magnetization and Exchange-Stiffness Constants of {{Fe}}–{{Al}}–{{Si}} Alloys at Finite Temperatures},
+ author = {Yamashita, Shogo and Sakuma, Akimasa},
+ date = {2024-07-02},
+ journaltitle = {Journal of Applied Physics},
+ shortjournal = {Journal of Applied Physics},
+ volume = {136},
+ number = {1},
+ pages = {013903},
+ issn = {0021-8979},
+ doi = {10.1063/5.0210430},
+ url = {https://doi.org/10.1063/5.0210430},
+ urldate = {2024-07-24},
+ abstract = {We investigated the magnetic properties of Sendust (Fe-Al-Si) alloys not only at 0 K but also at finite temperatures by means of the first-principles calculations assuming A2, B2, and D 0 3 structures. We confirmed that the itinerant characteristics of 3 d electrons of Fe are not negligible and a significantly small exchange stiffness constant exists at zero temperature in a B2 structure. However, the calculated Curie temperatures are in the same order for all structures; this indicates that the Curie temperature cannot be determined only by the exchange interactions at zero temperature in itinerant electron systems. Temperature dependence of the exchange interaction, namely, spin configuration dependence, also might be important for determining it. In addition, this property might also be related to the unique behavior of the temperature dependence of the exchange stiffness constant for the B2 structure, which does not decrease monotonically as temperatures increase, contrary to the behavior expected from the Heisenberg model. In addition, we investigated composition dependence on the exchange stiffness constant at zero temperature and confirmed that the substitution of Si with Al could improve the amplitude of the exchange stiffness constant at zero temperature for all structures.},
+ keywords = {/unread,FZJ,magnetization,PGI,PGI-1/IAS-1,PGI-1/IAS-1 guests},
+ file = {/Users/wasmer/Nextcloud/Zotero/Yamashita_Sakuma_2024_Magnetization and exchange-stiffness constants of Fe–Al–Si alloys at finite.pdf}
+}
+
@inproceedings{yamashitaTheoreticalInvestigationElectronic2022,
title = {Theoretical Investigation of Electronic Structure and Orbital Moment of the {{Sm}} Ions in {{SmFe12}} Using Generalized Gradient Approximation {{Theoretical Investigation}} of {{Electronic Structure}} and {{Orbital Moment}} of the {{Sm Ions}} in {{SmFe12}} Using {{Generalized Gradient Approximation}} +{{U MethodU}}\$ Method},
booktitle = {2022 {{Joint MMM-Intermag Conference}} ({{INTERMAG}})},
@@ -17840,7 +19440,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
url = {https://link.aps.org/doi/10.1103/PhysRevB.107.014407},
urldate = {2023-03-22},
abstract = {We present a systematic method to automatically generate symmetry-adapted magnetic structures for a given crystal structure and general propagation vector k as an efficient approach of the analysis of complex modulated magnetic structures. The method is developed as an extension of the generation scheme based on the multipole expansion, which was demonstrated only for the propagation vector k=0 [M.-T. Suzuki et al., Phys. Rev. B 99, 174407 (2019)]. The symmetry-adapted magnetic structures characterized with an ordering vector k are obtained by mapping the multipole magnetic alignments on a virtual cluster to the periodic crystal structure with the phase factor for the wave vector k. This method provides all magnetic bases compatible with irreducible representations under a k group for a given crystal structure and wave vector k. Multiple-k magnetic structures are derived from a superposition of single-k magnetic bases related to the space group symmetry. We apply the scheme to deduce the magnetic structures of α-Mn and CoM3S6 (M=Nb, Ta), in which the large anomalous Hall effect has recently been observed in antiferromagnetic phases, and identify the magnetic structures inducing anomalous Hall effect without net magnetization. The physical phenomena originating from emergent multipoles in the ordered phases are also discussed based on the Landau theory.},
- keywords = {\_tablet,AML,descriptors,feature engineering,Ferromagnetism,invariance,magnetism,ML,rec-by-kipp,spin-dependent},
+ keywords = {AML,descriptors,feature engineering,Ferromagnetism,invariance,magnetism,ML,rec-by-kipp,spin-dependent},
file = {/Users/wasmer/Zotero/storage/G9EKXIZ4/Yanagi et al_2023_Generation of modulated magnetic structures based on cluster multipole expansion.pdf;/Users/wasmer/Zotero/storage/ILBXKEJP/PhysRevB.107.html}
}
@@ -17883,6 +19483,23 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
file = {/home/johannes/Nextcloud/Zotero/false;/Users/wasmer/Zotero/storage/2L5JFJN8/5.html}
}
+@online{yangMatterSimDeepLearning2024,
+ title = {{{MatterSim}}: {{A Deep Learning Atomistic Model Across Elements}}, {{Temperatures}} and {{Pressures}}},
+ shorttitle = {{{MatterSim}}},
+ author = {Yang, Han and Hu, Chenxi and Zhou, Yichi and Liu, Xixian and Shi, Yu and Li, Jielan and Li, Guanzhi and Chen, Zekun and Chen, Shuizhou and Zeni, Claudio and Horton, Matthew and Pinsler, Robert and Fowler, Andrew and Zügner, Daniel and Xie, Tian and Smith, Jake and Sun, Lixin and Wang, Qian and Kong, Lingyu and Liu, Chang and Hao, Hongxia and Lu, Ziheng},
+ date = {2024-05-10},
+ eprint = {2405.04967},
+ eprinttype = {arXiv},
+ eprintclass = {cond-mat},
+ doi = {10.48550/arXiv.2405.04967},
+ url = {http://arxiv.org/abs/2405.04967},
+ urldate = {2024-06-26},
+ abstract = {Accurate and fast prediction of materials properties is central to the digital transformation of materials design. However, the vast design space and diverse operating conditions pose significant challenges for accurately modeling arbitrary material candidates and forecasting their properties. We present MatterSim, a deep learning model actively learned from large-scale first-principles computations, for efficient atomistic simulations at first-principles level and accurate prediction of broad material properties across the periodic table, spanning temperatures from 0 to 5000 K and pressures up to 1000 GPa. Out-of-the-box, the model serves as a machine learning force field, and shows remarkable capabilities not only in predicting ground-state material structures and energetics, but also in simulating their behavior under realistic temperatures and pressures, signifying an up to ten-fold enhancement in precision compared to the prior best-in-class. This enables MatterSim to compute materials' lattice dynamics, mechanical and thermodynamic properties, and beyond, to an accuracy comparable with first-principles methods. Specifically, MatterSim predicts Gibbs free energies for a wide range of inorganic solids with near-first-principles accuracy and achieves a 15 meV/atom resolution for temperatures up to 1000K compared with experiments. This opens an opportunity to predict experimental phase diagrams of materials at minimal computational cost. Moreover, MatterSim also serves as a platform for continuous learning and customization by integrating domain-specific data. The model can be fine-tuned for atomistic simulations at a desired level of theory or for direct structure-to-property predictions, achieving high data efficiency with a reduction in data requirements by up to 97\%.},
+ pubstate = {prepublished},
+ keywords = {/unread,active learning,Alexandria database,AML,fine-tuning,Graphormer,large models,M3GNet,MACE-MP-0,MatBench,materials discovery,materials project,MD,Microsoft Research,ML,MLFF,MLP,model comparison,out-of-distribution,PBE,PhononDB,Phonopy,prediction from structure,prediction of bandgap,prediction of formation energy,prediction of free energy,prediction of mechanical properties,prediction of phonon dispersion,property prediction,surrogate model,uncertainty quantification,universal potential,VASP,zero-shot MD},
+ file = {/Users/wasmer/Nextcloud/Zotero/Yang et al_2024_MatterSim.pdf;/Users/wasmer/Zotero/storage/LPW43T45/2405.html}
+}
+
@article{yaoTensorMol0ModelChemistry2018,
title = {The {{TensorMol-0}}.1 Model Chemistry: A Neural Network Augmented with Long-Range Physics},
shorttitle = {The {{TensorMol-0}}.1 Model Chemistry},
@@ -17906,14 +19523,14 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Yuan, Zilong and Xu, Zhiming and Li, He and Cheng, Xinle and Tao, Honggeng and Tang, Zechen and Zhou, Zhiyuan and Duan, Wenhui and Xu, Yong},
date = {2024-02-07},
eprint = {2402.04864},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2402.04864},
url = {http://arxiv.org/abs/2402.04864},
urldate = {2024-04-18},
abstract = {Neural network force fields have significantly advanced ab initio atomistic simulations across diverse fields. However, their application in the realm of magnetic materials is still in its early stage due to challenges posed by the subtle magnetic energy landscape and the difficulty of obtaining training data. Here we introduce a data-efficient neural network architecture to represent density functional theory total energy, atomic forces, and magnetic forces as functions of atomic and magnetic structures. Our approach incorporates the principle of equivariance under the three-dimensional Euclidean group into the neural network model. Through systematic experiments on various systems, including monolayer magnets, curved nanotube magnets, and moir\textbackslash 'e-twisted bilayer magnets of \$\textbackslash text\{CrI\}\_\{3\}\$, we showcase the method's high efficiency and accuracy, as well as exceptional generalization ability. The work creates opportunities for exploring magnetic phenomena in large-scale materials systems.},
- pubstate = {preprint},
- keywords = {\_tablet,2D material,AML,CNT,constrained DFT,DeepH,Landau-Lifshits-Gilbert equation,magnetism,magnon dispersion,magnons,ML,MLP,non-collinear,prediction from magnetic configuration,prediction from structure,prediction of energy,prediction of forces,skyrmions,spin dynamics,spin wave,twisted bilayer,VASP},
+ pubstate = {prepublished},
+ keywords = {2D material,AML,CNT,constrained DFT,DeepH,Landau-Lifshits-Gilbert equation,magnetism,magnon dispersion,magnons,ML,MLP,non-collinear,prediction from magnetic configuration,prediction from structure,prediction of energy,prediction of forces,skyrmions,spin dynamics,spin wave,twisted bilayer,VASP},
file = {/Users/wasmer/Nextcloud/Zotero/Yuan et al_2024_Equivariant Neural Network Force Fields for Magnetic Materials.pdf;/Users/wasmer/Zotero/storage/FB8XK7MY/2402.html}
}
@@ -17922,13 +19539,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Yu, Hongyu and Hong, Liangliang and Chen, Shiyou and Gong, Xingao and Xiang, Hongjun},
date = {2022-11-29},
eprint = {2211.16684},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2211.16684},
url = {http://arxiv.org/abs/2211.16684},
urldate = {2023-10-13},
abstract = {Machine Learning (ML) interatomic models and potentials have been widely employed in simulations of materials. Long-range interactions often dominate in some ionic systems whose dynamics behavior is significantly influenced. However, the long-range effect such as Coulomb and Van der Wales potential is not considered in most ML interatomic potentials. To address this issue, we put forward a method that can take long-range effects into account for most ML local interatomic models with the reciprocal space neural network. The structure information in real space is firstly transformed into reciprocal space and then encoded into a reciprocal space potential or a global descriptor with full atomic interactions. The reciprocal space potential and descriptor keep full invariance of Euclidean symmetry and choice of the cell. Benefiting from the reciprocal-space information, ML interatomic models can be extended to describe the long-range potential including not only Coulomb but any other long-range interaction. A model NaCl system considering Coulomb interaction and the GaxNy system with defects are applied to illustrate the advantage of our approach. At the same time, our approach helps to improve the prediction accuracy of some global properties such as the band gap where the full atomic interaction beyond local atomic environments plays a very important role. In summary, our work has expanded the ability of current ML interatomic models and potentials when dealing with the long-range effect, hence paving a new way for accurate prediction of global properties and large-scale dynamic simulations of systems with defects.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,CGCNN,charge transfer,descriptors,DimeNet,E(3),invariance,long-range interaction,materials,ML,MLP,PAiNN,reciprocal space,reciprocal space descriptor,vdW},
file = {/Users/wasmer/Nextcloud/Zotero/Yu et al_2022_Capturing long-range interaction with reciprocal space neural network.pdf;/Users/wasmer/Zotero/storage/KHNJJ4VF/2211.html}
}
@@ -17956,13 +19573,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Yu, Haiyang and Xu, Zhao and Qian, Xiaofeng and Qian, Xiaoning and Ji, Shuiwang},
date = {2023-06-07},
eprint = {2306.04922},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2306.04922},
url = {http://arxiv.org/abs/2306.04922},
urldate = {2023-10-13},
abstract = {We consider the prediction of the Hamiltonian matrix, which finds use in quantum chemistry and condensed matter physics. Efficiency and equivariance are two important, but conflicting factors. In this work, we propose a SE(3)-equivariant network, named QHNet, that achieves efficiency and equivariance. Our key advance lies at the innovative design of QHNet architecture, which not only obeys the underlying symmetries, but also enables the reduction of number of tensor products by 92\textbackslash\%. In addition, QHNet prevents the exponential growth of channel dimension when more atom types are involved. We perform experiments on MD17 datasets, including four molecular systems. Experimental results show that our QHNet can achieve comparable performance to the state of the art methods at a significantly faster speed. Besides, our QHNet consumes 50\textbackslash\% less memory due to its streamlined architecture. Our code is publicly available as part of the AIRS library (\textbackslash url\{https://github.com/divelab/AIRS\}).},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ablation study,AML,chemical species scaling problem,DFT,e3nn,emulator,equivariant,GNN,hybrid AI/simulation,library,ML,ML-DFT,ML-ESM,molecules,MPNN,PBE,PhiSNet,prediction of Hamiltonian matrix,PySCF,QHNet,quantum tensors,SE(3),with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Yu et al_2023_Efficient and Equivariant Graph Networks for Predicting Quantum Hamiltonian.pdf;/Users/wasmer/Zotero/storage/8MAS3EXH/2306.html}
}
@@ -17990,14 +19607,14 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Yu, Hongyu and Liu, Boyu and Zhong, Yang and Hong, Liangliang and Ji, Junyi and Xu, Changsong and Gong, Xingao and Xiang, Hongjun},
date = {2024-01-08},
eprint = {2211.11403},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2211.11403},
url = {http://arxiv.org/abs/2211.11403},
urldate = {2024-04-12},
abstract = {This study introduces time-reversal E(3)-equivariant neural network and SpinGNN++ framework for constructing a comprehensive interatomic potential for magnetic systems, encompassing spin-orbit coupling and noncollinear magnetic moments. SpinGNN++ integrates multitask spin equivariant neural network with explicit spin-lattice terms, including Heisenberg, Dzyaloshinskii-Moriya, Kitaev, single-ion anisotropy, and biquadratic interactions, and employs time-reversal equivariant neural network to learn high-order spin-lattice interactions using time-reversal E(3)-equivariant convolutions. To validate SpinGNN++, a complex magnetic model dataset is introduced as a benchmark and employed to demonstrate its capabilities. SpinGNN++ provides accurate descriptions of the complex spin-lattice coupling in monolayer CrI\$\_3\$ and CrTe\$\_2\$, achieving sub-meV errors. Importantly, it facilitates large-scale parallel spin-lattice dynamics, thereby enabling the exploration of associated properties, including the magnetic ground state and phase transition. Remarkably, SpinGNN++ identifies a new ferrimagnetic state as the ground magnetic state for monolayer CrTe2, thereby enriching its phase diagram and providing deeper insights into the distinct magnetic signals observed in various experiments.},
- pubstate = {preprint},
- keywords = {\_tablet,Allegro,AML,GNN,ML,ML-DFT,ML-ESM,MLP,MPNN,NequIP,non-collinear,prediction of Jij,SOC,spin-dependent,TRS,with-code},
+ pubstate = {prepublished},
+ keywords = {Allegro,AML,GNN,ML,ML-DFT,ML-ESM,MLP,MPNN,NequIP,non-collinear,prediction of Jij,SOC,spin-dependent,TRS,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Yu et al_2024_General time-reversal equivariant neural network potential for magnetic.pdf;/Users/wasmer/Zotero/storage/XV2X47EC/2211.html}
}
@@ -18007,13 +19624,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Yu, Haiyang and Liu, Meng and Luo, Youzhi and Strasser, Alex and Qian, Xiaofeng and Qian, Xiaoning and Ji, Shuiwang},
date = {2023-06-15},
eprint = {2306.09549},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2306.09549},
url = {http://arxiv.org/abs/2306.09549},
urldate = {2023-10-13},
abstract = {Supervised machine learning approaches have been increasingly used in accelerating electronic structure prediction as surrogates of first-principle computational methods, such as density functional theory (DFT). While numerous quantum chemistry datasets focus on chemical properties and atomic forces, the ability to achieve accurate and efficient prediction of the Hamiltonian matrix is highly desired, as it is the most important and fundamental physical quantity that determines the quantum states of physical systems and chemical properties. In this work, we generate a new Quantum Hamiltonian dataset, named as QH9, to provide precise Hamiltonian matrices for 2,399 molecular dynamics trajectories and 130,831 stable molecular geometries, based on the QM9 dataset. By designing benchmark tasks with various molecules, we show that current machine learning models have the capacity to predict Hamiltonian matrices for arbitrary molecules. Both the QH9 dataset and the baseline models are provided to the community through an open-source benchmark, which can be highly valuable for developing machine learning methods and accelerating molecular and materials design for scientific and technological applications. Our benchmark is publicly available at https://github.com/divelab/AIRS/tree/main/OpenDFT/QHBench.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,benchmark dataset,benchmarking,Database,dataset,DeepH,DFT speedup,DFT speedup with ML,equivariant,ML,ML-DFT,ML-ESM,molecules,out-of-distribution,prediction of energy,prediction of Hamiltonian matrix,prediction of wavefunction,QHNet,QM9,SchNOrb,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Yu et al_2023_QH9.pdf;/Users/wasmer/Zotero/storage/LHYKLK48/2306.html}
}
@@ -18027,8 +19644,8 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
urldate = {2023-06-12},
abstract = {The development of machine learning interatomic potentials has immensely contributed to the accuracy of simulations of molecules and crystals. However, creating interatomic potentials for magnetic systems that account for both magnetic moments and structural degrees of freedom remains a challe...},
langid = {english},
- pubstate = {preprint},
- keywords = {\_tablet,Allegro,AML,collinear,DimeNet,equivariant,GNN,HDNNP,heat transport,Heisenberg model,Jij,LAMMPS,LAMMPS SPIN,magnetism,MD,mHDNNP,ML,MLP,mMTP,MTP,multiferroic,non-collinear,original publication,PES,prediction of Jij,prediction of magnetic ground state,skyrmions,spin dynamics,Spin-Allegro,spin-dependent,Spin-Dimenet,spin-lattice coupling,SpinGNN,with-code},
+ pubstate = {prepublished},
+ keywords = {Allegro,AML,collinear,DimeNet,equivariant,GNN,HDNNP,heat transport,Heisenberg model,Jij,LAMMPS,LAMMPS SPIN,magnetism,MD,mHDNNP,ML,MLP,mMTP,MTP,multiferroic,non-collinear,original publication,PES,prediction of Jij,prediction of magnetic ground state,skyrmions,spin dynamics,Spin-Allegro,spin-dependent,Spin-Dimenet,spin-lattice coupling,SpinGNN,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Yu et al_2023_Spin-Dependent Graph Neural Network Potential for Magnetic Materials.pdf}
}
@@ -18037,15 +19654,15 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Yu, Hongyu and Zhong, Yang and Hong, Liangliang and Xu, Changsong and Ren, Wei and Gong, Xingao and Xiang, Hongjun},
date = {2023-04-20},
eprint = {2203.02853},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2203.02853},
url = {http://arxiv.org/abs/2203.02853},
urldate = {2023-10-14},
abstract = {The development of machine learning interatomic potentials has immensely contributed to the accuracy of simulations of molecules and crystals. However, creating interatomic potentials for magnetic systems that account for both magnetic moments and structural degrees of freedom remains a challenge. This work introduces SpinGNN, a spin-dependent interatomic potential approach that employs the graph neural network (GNN) to describe magnetic systems. SpinGNN consists of two types of edge GNNs: Heisenberg edge GNN (HEGNN) and spin-distance edge GNN (SEGNN). HEGNN is tailored to capture Heisenberg-type spin-lattice interactions, while SEGNN accurately models multi-body and high-order spin-lattice coupling. The effectiveness of SpinGNN is demonstrated by its exceptional precision in fitting a high-order spin Hamiltonian and two complex spin-lattice Hamiltonians with great precision. Furthermore, it successfully models the subtle spin-lattice coupling in BiFeO3 and performs large-scale spin-lattice dynamics simulations, predicting its antiferromagnetic ground state, magnetic phase transition, and domain wall energy landscape with high accuracy. Our study broadens the scope of graph neural network potentials to magnetic systems, serving as a foundation for carrying out large-scale spin-lattice dynamic simulations of such systems.},
- pubstate = {preprint},
- keywords = {\_tablet,Allegro,AML,collinear,DimeNet,equivariant,GNN,HDNNP,heat transport,Heisenberg model,Jij,LAMMPS,LAMMPS SPIN,magnetism,MD,mHDNNP,ML,MLP,mMTP,MTP,multiferroic,non-collinear,original publication,PES,prediction of Jij,prediction of magnetic ground state,skyrmions,spin dynamics,Spin-Allegro,spin-dependent,Spin-Dimenet,spin-lattice coupling,SpinGNN,with-code},
- file = {/Users/wasmer/Nextcloud/Zotero/false;/Users/wasmer/Zotero/storage/B3HN773D/2203.html}
+ pubstate = {prepublished},
+ keywords = {Allegro,AML,collinear,DimeNet,equivariant,GNN,HDNNP,heat transport,Heisenberg model,Jij,LAMMPS,LAMMPS SPIN,magnetism,MD,mHDNNP,ML,MLP,mMTP,MTP,multiferroic,non-collinear,original publication,PES,prediction of Jij,prediction of magnetic ground state,skyrmions,spin dynamics,Spin-Allegro,spin-dependent,Spin-Dimenet,spin-lattice coupling,SpinGNN,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Yu et al_2023_Spin-Dependent Graph Neural Network Potential for Magnetic Materials-arxiv.pdf;/Users/wasmer/Zotero/storage/B3HN773D/2203.html}
}
@book{zabloudilElectronScatteringSolid2005,
@@ -18088,7 +19705,27 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
file = {/Users/wasmer/Nextcloud/Zotero/Zaharia et al_2018_Accelerating the Machine Learning Lifecycle with MLflow.pdf}
}
-@article{zaverkinFastSampleEfficientInteratomic2021,
+@article{zaverkinExploringChemicalConformational2022,
+ title = {Exploring Chemical and Conformational Spaces by Batch Mode Deep Active Learning},
+ author = {Zaverkin, Viktor and Holzmüller, David and Steinwart, Ingo and Kästner, Johannes},
+ date = {2022-10-10},
+ journaltitle = {Digital Discovery},
+ shortjournal = {Digital Discovery},
+ volume = {1},
+ number = {5},
+ pages = {605--620},
+ publisher = {RSC},
+ issn = {2635-098X},
+ doi = {10.1039/D2DD00034B},
+ url = {https://pubs.rsc.org/en/content/articlelanding/2022/dd/d2dd00034b},
+ urldate = {2024-06-15},
+ abstract = {The development of machine-learned interatomic potentials requires generating sufficiently expressive atomistic data sets. Active learning algorithms select data points on which labels, i.e., energies and forces, are calculated for inclusion in the training set. However, for batch mode active learning, i.e., when multiple data points are selected at once, conventional active learning algorithms can perform poorly. Therefore, we investigate algorithms specifically designed for this setting and show that they can outperform conventional algorithms. We investigate selection based on the informativeness, diversity, and representativeness of the resulting training set. We propose using gradient features specific to atomistic neural networks to evaluate the informativeness of queried samples, including several approximations allowing for their efficient evaluation. To avoid selecting similar structures, we present several methods that enforce the diversity and representativeness of the selected batch. Finally, we apply the proposed approaches to several molecular and periodic bulk benchmark systems and argue that they can be used to generate highly informative atomistic data sets by running any atomistic simulation.},
+ langid = {english},
+ keywords = {/unread,active learning,AML,descriptors,GM-NN,GPR,GTO basis,kernel methods,library,ML,MLP,similarity kernel,similarity measure,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Zaverkin et al_2022_Exploring chemical and conformational spaces by batch mode deep active learning.pdf;/Users/wasmer/Zotero/storage/SM4Y4NW5/Zaverkin et al. - 2022 - Exploring chemical and conformational spaces by ba.pdf}
+}
+
+@article{zaverkinFastSampleEfficientInteratomic2021a,
title = {Fast and {{Sample-Efficient Interatomic Neural Network Potentials}} for {{Molecules}} and {{Materials Based}} on {{Gaussian Moments}}},
author = {Zaverkin, Viktor and Holzmüller, David and Steinwart, Ingo and Kästner, Johannes},
date = {2021-10-12},
@@ -18101,9 +19738,49 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
issn = {1549-9618},
doi = {10.1021/acs.jctc.1c00527},
url = {https://doi.org/10.1021/acs.jctc.1c00527},
- urldate = {2022-01-02},
+ urldate = {2024-06-15},
abstract = {Artificial neural networks (NNs) are one of the most frequently used machine learning approaches to construct interatomic potentials and enable efficient large-scale atomistic simulations with almost ab initio accuracy. However, the simultaneous training of NNs on energies and forces, which are a prerequisite for, e.g., molecular dynamics simulations, can be demanding. In this work, we present an improved NN architecture based on the previous GM-NN model [Zaverkin V.; Kästner, J. J. Chem. Theory Comput. 2020, 16, 5410−5421], which shows an improved prediction accuracy and considerably reduced training times. Moreover, we extend the applicability of Gaussian moment-based interatomic potentials to periodic systems and demonstrate the overall excellent transferability and robustness of the respective models. The fast training by the improved methodology is a prerequisite for training-heavy workflows such as active learning or learning-on-the-fly.},
- file = {/Users/wasmer/Nextcloud/Zotero/Zaverkin et al_2021_Fast and Sample-Efficient Interatomic Neural Network Potentials for Molecules.pdf}
+ keywords = {/unread,AML,descriptors,FCHL,GM-NN,GTO basis,library,MD,MD17,ML,MLP,model comparison,PAiNN,PhysNet,QM9,quaternary systems,SchNet,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Zaverkin et al_2021_Fast and Sample-Efficient Interatomic Neural Network Potentials for Molecules2.pdf}
+}
+
+@article{zaverkinGaussianMomentsPhysically2020,
+ title = {Gaussian {{Moments}} as {{Physically Inspired Molecular Descriptors}} for {{Accurate}} and {{Scalable Machine Learning Potentials}}},
+ author = {Zaverkin, V. and Kästner, J.},
+ date = {2020-08-11},
+ journaltitle = {Journal of Chemical Theory and Computation},
+ shortjournal = {J. Chem. Theory Comput.},
+ volume = {16},
+ number = {8},
+ pages = {5410--5421},
+ publisher = {American Chemical Society},
+ issn = {1549-9618},
+ doi = {10.1021/acs.jctc.0c00347},
+ url = {https://doi.org/10.1021/acs.jctc.0c00347},
+ urldate = {2024-06-15},
+ abstract = {Machine learning techniques allow a direct mapping of atomic positions and nuclear charges to the potential energy surface with almost ab initio accuracy and the computational efficiency of empirical potentials. In this work, we propose a machine learning method for constructing high-dimensional potential energy surfaces based on feed-forward neural networks. As input to the neural network, we propose an extendable invariant local molecular descriptor constructed from geometric moments. Their formulation via pairwise distance vectors and tensor contractions allows a very efficient implementation on graphical processing units (GPUs). The atomic species is encoded in the molecular descriptor, which allows the restriction to one neural network for the training of all atomic species in the data set. We demonstrate that the accuracy of the developed approach in representing both chemical and configurational spaces is comparable to the one of several established machine learning models. Due to its high accuracy and efficiency, the proposed machine-learned potentials can be used for any further tasks, for example, the optimization of molecular geometries, the calculation of rate constants, or molecular dynamics.},
+ keywords = {/unread,AML,descriptors,GM-NN,GTO basis,library,MD17,ML,MLP,original publication,PhysNet,SchNet,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Zaverkin_Kästner_2020_Gaussian Moments as Physically Inspired Molecular Descriptors for Accurate and.pdf}
+}
+
+@article{zaverkinUncertaintybiasedMolecularDynamics2024,
+ title = {Uncertainty-Biased Molecular Dynamics for Learning Uniformly Accurate Interatomic Potentials},
+ author = {Zaverkin, Viktor and Holzmüller, David and Christiansen, Henrik and Errica, Federico and Alesiani, Francesco and Takamoto, Makoto and Niepert, Mathias and Kästner, Johannes},
+ date = {2024-04-29},
+ journaltitle = {npj Computational Materials},
+ shortjournal = {npj Comput Mater},
+ volume = {10},
+ number = {1},
+ pages = {1--18},
+ publisher = {Nature Publishing Group},
+ issn = {2057-3960},
+ doi = {10.1038/s41524-024-01254-1},
+ url = {https://www.nature.com/articles/s41524-024-01254-1},
+ urldate = {2024-06-15},
+ abstract = {Efficiently creating a concise but comprehensive data set for training machine-learned interatomic potentials (MLIPs) is an under-explored problem. Active learning, which uses biased or unbiased molecular dynamics (MD) to generate candidate pools, aims to address this objective. Existing biased and unbiased MD-simulation methods, however, are prone to miss either rare events or extrapolative regions—areas of the configurational space where unreliable predictions are made. This work demonstrates that MD, when biased by the MLIP’s energy uncertainty, simultaneously captures extrapolative regions and rare events, which is crucial for developing uniformly accurate MLIPs. Furthermore, exploiting automatic differentiation, we enhance bias-forces-driven MD with the concept of bias stress. We employ calibrated gradient-based uncertainties to yield MLIPs with similar or, sometimes, better accuracy than ensemble-based methods at a lower computational cost. Finally, we apply uncertainty-biased MD to alanine dipeptide and MIL-53(Al), generating MLIPs that represent both configurational spaces more accurately than models trained with conventional MD.},
+ langid = {english},
+ keywords = {/unread,active learning,active learning protocol,AML,descriptors,GM-NN,GTO basis,library,MD,ML,MLP,uncertainty quantification,with-code},
+ file = {/Users/wasmer/Nextcloud/Zotero/Zaverkin et al_2024_Uncertainty-biased molecular dynamics for learning uniformly accurate.pdf}
}
@online{zeerPromotingPbasedHall2023,
@@ -18111,13 +19788,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zeer, Mahmoud and Go, Dongwook and Schmitz, Peter and Saunderson, Tom G. and Wang, Hao and Ghabboun, Jamal and Blügel, Stefan and Wulfhekel, Wulf and Mokrousov, Yuriy},
date = {2023-08-16},
eprint = {2308.08207},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2308.08207},
url = {http://arxiv.org/abs/2308.08207},
urldate = {2023-10-04},
abstract = {We conduct a first-principles study of Hall effects in rare-earth dichalcogenides, focusing on monolayers of the H-phase EuX\$\_2\$ and GdX\$\_2\$, where X = S, Se, and Te. Our predictions reveal that all EuX\$\_2\$ and GdX\$\_2\$ systems exhibit high magnetic moments and wide bandgaps. We observe that while in case of EuX\$\_2\$ the \$p\$ and \$f\$ states hybridize directly below the Fermi energy, the absence of \$f\$ and \$d\$ states of Gd at the Fermi energy results in \$p\$-like spin-polarized electronic structure of GdX\$\_2\$, which mediates \$p\$-based magnetotransport. Notably, these systems display significant anomalous, spin, and orbital Hall conductivities. We find that in GdX\$\_2\$ the strength of correlations controls the relative position of \$p\$, \$d\$ and \$f\$-states and their hybridization which has a crucial impact on \$p\$-state polarization and the anomalous Hall effect, but not the spin and orbital Hall effect. Moreover, we find that the application of strain can significantly modify the electronic structure of the monolayers, resulting in quantized charge, spin and orbital transport in GdTe\$\_2\$ via a strain-mediated orbital inversion mechanism taking place at the Fermi energy. Our findings suggest that rare-earth dichalcogenides hold promise as a platform for topological spintronics and orbitronics.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {chalcogenides,DFT,DFT+U,FLEUR,FZJ,Hall effect,Hall OHE,PGI,PGI-1/IAS-1,physics,quantum materials,rare earths,SOC,Wannier},
file = {/Users/wasmer/Nextcloud/Zotero/Zeer et al_2023_Promoting $p$-based Hall effects by $p$-$d$-$f$ hybridization in Gd-based.pdf;/Users/wasmer/Zotero/storage/N7TZNGEI/2308.html}
}
@@ -18155,10 +19832,28 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
abstract = {Density-functional theory (DFT) is considered the Standard Model of solid-state physics. The state-of-the-art approximations to DFT, the local-density approximation (LDA) or its simple extensions, fail, however, even qualitatively, for strongly-correlated systems. When correlations are strong, electrons become entangled and novel properties emerge. Mott-transitions, Kondo- and heavy-fermion behavior, non-conventional superconductivity and orbital-order are just some examples of this emergent behavior. The realistic description of emergent properties is one of the grand-challenges of modern condensed-matter physics. To understand this physics beyond the Standard Model, nonperturbative many-body techniques are essential. Still, DFT-based methods are needed to devise materials-specific Hamiltonians for strong correlations. Mastering these novel techniques requires a vast background, ranging from DFT to model building and many-body physics. The aim of this school is to introduce advanced graduate students and up to the modern methods for modeling emergent properties of correlated electrons and to explore the relation of electron correlations with quantum entanglement and concepts from quantum information. A school of this size and scope requires support and help from many sources. We are very grateful for all the financial and practical support we have received. The Institute for Advanced Simulation and the German Research School for Simulation Sciences at the Forschungszentrum Jülich provided the funding and were vital for the organization of the school and the production of this book. The DFG Forschergruppe FOR1346 offered travel grants for students and the Institute for Complex Adaptive Matter (ICAM) travel support for international speakers and participants. The nature of a school makes it desirable to have the lecture-notes available already during the lectures. In this way the participants get the chance to work through the lectures thoroughly while they are given. We are therefore extremely grateful to the lecturers that, despite a tight schedule, provided their manuscripts in time for the production of this book. We are confident that the lecture notes collected here will not only serve the participants of the school but will also be useful for other students entering the exciting field of strongly correlated materials. We thank Mrs. H. Lexis of the Forschungszentrum Jülich Verlag and Mr. D. Laufenberg of the Graphische Betriebe for providing their expert support in producing the present volume on a tight schedule and for making even seemingly impossible requests possible. We heartily thank our students and postdocs that helped in proofreading the manuscripts, often on short notice: Carmine Autieri, Fabio Baruffa, Michael Baumgärtel, Monica Bugeanu, Andreas Flesch, Evgeny Gorelov, Amin Kiani Sheikhabadi, Joaquin Miranda, German Ulm, and Guoren Zhang. Finally, our special thanks go to Dipl.-Ing. R. Hölzle for his invaluable advice on all questions concerning the organization of such a school and to Mrs. L. Snyders and Mrs. E. George for expertly handling all practical issues. Pavarini, Eva; Koch, Erik; Anders, Frithjof; Jarrell, Mark (Eds. )},
isbn = {9783893367962},
langid = {english},
- keywords = {\_tablet,CPA,KKR,PGI-1/IAS-1,VCA},
+ keywords = {CPA,KKR,PGI-1/IAS-1,VCA},
file = {/Users/wasmer/Nextcloud/Zotero/Zeller_2012_Correlated electrons.pdf;/Users/wasmer/Zotero/storage/BKBRXSWN/136393.html}
}
+@article{zellerLloydFormulaMultiplescattering2005,
+ title = {Lloyd’s Formula in Multiple-Scattering Calculations with Finite Temperature},
+ author = {Zeller, Rudolf},
+ date = {2005-08},
+ journaltitle = {Journal of Physics: Condensed Matter},
+ shortjournal = {J. Phys.: Condens. Matter},
+ volume = {17},
+ number = {35},
+ pages = {5367},
+ issn = {0953-8984},
+ doi = {10.1088/0953-8984/17/35/005},
+ url = {https://dx.doi.org/10.1088/0953-8984/17/35/005},
+ urldate = {2024-07-05},
+ abstract = {Lloyd’s formula is an elegant tool to calculate the number of states directly from the imaginary part of the logarithm of the Korringa–Kohn–Rostoker (KKR) determinant. It is shown how this formula can be used at finite electronic temperatures and how the difficult problem to determine the physically significant correct phase of the complex logarithm can be circumvented by working with the single-valued real part of the logarithm. The approach is based on contour integrations in the complex energy plane and exploits the analytical properties of the KKR Green function and the Fermi–Dirac function. It leads to rather accurate results, which is illustrated by a local-density functional calculation of the temperature dependence of the intrinsic Fermi level in zinc-blende GaN.},
+ langid = {english},
+ keywords = {/unread}
+}
+
@book{zengQuantumInformationMeets2019,
title = {Quantum {{Information Meets Quantum Matter}}: {{From Quantum Entanglement}} to {{Topological Phases}} of {{Many-Body Systems}}},
shorttitle = {Quantum {{Information Meets Quantum Matter}}},
@@ -18191,7 +19886,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
url = {https://aip.scitation.org/doi/10.1063/5.0052961},
urldate = {2022-05-11},
abstract = {We probe the accuracy of linear ridge regression employing a three-body local density representation derived from the atomic cluster expansion. We benchmark the accuracy of this framework in the prediction of formation energies and atomic forces in molecules and solids. We find that such a simple regression framework performs on par with state-of-the-art machine learning methods which are, in most cases, more complex and more computationally demanding. Subsequently, we look for ways to sparsify the descriptor and further improve the computational efficiency of the method. To this aim, we use both principal component analysis and least absolute shrinkage operator regression for energy fitting on six single-element datasets. Both methods highlight the possibility of constructing a descriptor that is four times smaller than the original with a similar or even improved accuracy. Furthermore, we find that the reduced descriptors share a sizable fraction of their features across the six independent datasets, hinting at the possibility of designing material-agnostic, optimally compressed, and accurate descriptors.},
- keywords = {\_tablet,ACE,descriptor dimred,descriptors,dimensionality reduction},
+ keywords = {ACE,descriptor dimred,descriptors,dimensionality reduction},
file = {/Users/wasmer/Nextcloud/Zotero/Zeni et al_2021_Compact atomic descriptors enable accurate predictions via linear models.pdf}
}
@@ -18200,13 +19895,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zenil, Hector and Tegnér, Jesper and Abrahão, Felipe S. and Lavin, Alexander and Kumar, Vipin and Frey, Jeremy G. and Weller, Adrian and Soldatova, Larisa and Bundy, Alan R. and Jennings, Nicholas R. and Takahashi, Koichi and Hunter, Lawrence and Dzeroski, Saso and Briggs, Andrew and Gregory, Frederick D. and Gomes, Carla P. and Williams, Christopher K. I. and Rowe, Jon and Evans, James and Kitano, Hiroaki and Tenenbaum, Joshua B. and King, Ross},
date = {2023-08-14},
eprint = {2307.07522},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2307.07522},
url = {http://arxiv.org/abs/2307.07522},
urldate = {2023-08-22},
abstract = {Recent advances in machine learning and AI, including Generative AI and LLMs, are disrupting technological innovation, product development, and society as a whole. AI's contribution to technology can come from multiple approaches that require access to large training data sets and clear performance evaluation criteria, ranging from pattern recognition and classification to generative models. Yet, AI has contributed less to fundamental science in part because large data sets of high-quality data for scientific practice and model discovery are more difficult to access. Generative AI, in general, and Large Language Models in particular, may represent an opportunity to augment and accelerate the scientific discovery of fundamental deep science with quantitative models. Here we explore and investigate aspects of an AI-driven, automated, closed-loop approach to scientific discovery, including self-driven hypothesis generation and open-ended autonomous exploration of the hypothesis space. Integrating AI-driven automation into the practice of science would mitigate current problems, including the replication of findings, systematic production of data, and ultimately democratisation of the scientific process. Realising these possibilities requires a vision for augmented AI coupled with a diversity of AI approaches able to deal with fundamental aspects of causality analysis and model discovery while enabling unbiased search across the space of putative explanations. These advances hold the promise to unleash AI's potential for searching and discovering the fundamental structure of our world beyond what human scientists have been able to achieve. Such a vision would push the boundaries of new fundamental science rather than automatize current workflows and instead open doors for technological innovation to tackle some of the greatest challenges facing humanity today.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,Pasteur \& ISI,todo-tagging},
file = {/Users/wasmer/Zotero/storage/RXPDN4KU/Zenil et al. - 2023 - The Future of Fundamental Science Led by Generativ.pdf;/Users/wasmer/Zotero/storage/FRVYJZDW/2307.html}
}
@@ -18217,14 +19912,14 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zeni, Claudio and Pinsler, Robert and Zügner, Daniel and Fowler, Andrew and Horton, Matthew and Fu, Xiang and Shysheya, Sasha and Crabbé, Jonathan and Sun, Lixin and Smith, Jake and Tomioka, Ryota and Xie, Tian},
date = {2023-12-06},
eprint = {2312.03687},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat},
doi = {10.48550/arXiv.2312.03687},
url = {http://arxiv.org/abs/2312.03687},
urldate = {2023-12-17},
abstract = {The design of functional materials with desired properties is essential in driving technological advances in areas like energy storage, catalysis, and carbon capture. Generative models provide a new paradigm for materials design by directly generating entirely novel materials given desired property constraints. Despite recent progress, current generative models have low success rate in proposing stable crystals, or can only satisfy a very limited set of property constraints. Here, we present MatterGen, a model that generates stable, diverse inorganic materials across the periodic table and can further be fine-tuned to steer the generation towards a broad range of property constraints. To enable this, we introduce a new diffusion-based generative process that produces crystalline structures by gradually refining atom types, coordinates, and the periodic lattice. We further introduce adapter modules to enable fine-tuning towards any given property constraints with a labeled dataset. Compared to prior generative models, structures produced by MatterGen are more than twice as likely to be novel and stable, and more than 15 times closer to the local energy minimum. After fine-tuning, MatterGen successfully generates stable, novel materials with desired chemistry, symmetry, as well as mechanical, electronic and magnetic properties. Finally, we demonstrate multi-property materials design capabilities by proposing structures that have both high magnetic density and a chemical composition with low supply-chain risk. We believe that the quality of generated materials and the breadth of MatterGen's capabilities represent a major advancement towards creating a universal generative model for materials design.},
- pubstate = {preprint},
- keywords = {/unread,AI4Science,AML,database generation,DFT,diffusion model,generative models,M3GNet,magnetic density,materials discovery,Microsoft Research,ML,n-ary alloys,structure prediction,symmetry,workflows},
+ pubstate = {prepublished},
+ keywords = {AI4Science,AML,database generation,DFT,diffusion model,generative models,M3GNet,magnetic density,materials discovery,Microsoft Research,ML,n-ary alloys,structure prediction,symmetry,workflows},
file = {/Users/wasmer/Nextcloud/Zotero/Zeni et al_2023_MatterGen.pdf;/Users/wasmer/Zotero/storage/9J3S7YXT/2312.html}
}
@@ -18243,7 +19938,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
urldate = {2022-07-08},
abstract = {The recently developed Deep Potential [Phys. Rev. Lett. 120 (2018) 143001 [27]] is a powerful method to represent general inter-atomic potentials using deep neural networks. The success of Deep Potential rests on the proper treatment of locality and symmetry properties of each component of the network. In this paper, we leverage its network structure to effectively represent the mapping from the atomic configuration to the electron density in Kohn-Sham density function theory (KS-DFT). By directly targeting at the self-consistent electron density, we demonstrate that the adapted network architecture, called the Deep Density, can effectively represent the self-consistent electron density as the linear combination of contributions from many local clusters. The network is constructed to satisfy the translation, rotation, and permutation symmetries, and is designed to be transferable to different system sizes. We demonstrate that using a relatively small number of training snapshots, with each snapshot containing a modest amount of data-points, Deep Density achieves excellent performance for one-dimensional insulating and metallic systems, as well as systems with mixed insulating and metallic characters. We also demonstrate its performance for real three-dimensional systems, including small organic molecules, as well as extended systems such as water (up to 512 molecules) and aluminum (up to 256 atoms).},
langid = {english},
- keywords = {\_tablet,DFT,ML,ML-DFT,ML-ESM,prediction of electron density},
+ keywords = {DFT,ML,ML-DFT,ML-ESM,prediction of electron density},
file = {/Users/wasmer/Nextcloud/Zotero/Zepeda-Núñez et al_2021_Deep Density.pdf;/Users/wasmer/Zotero/storage/TJJ4NCEI/S0021999121004186.html}
}
@@ -18252,36 +19947,32 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhang, Xuan and Wang, Limei and Helwig, Jacob and Luo, Youzhi and Fu, Cong and Xie, Yaochen and Liu, Meng and Lin, Yuchao and Xu, Zhao and Yan, Keqiang and Adams, Keir and Weiler, Maurice and Li, Xiner and Fu, Tianfan and Wang, Yucheng and Yu, Haiyang and Xie, YuQing and Fu, Xiang and Strasser, Alex and Xu, Shenglong and Liu, Yi and Du, Yuanqi and Saxton, Alexandra and Ling, Hongyi and Lawrence, Hannah and Stärk, Hannes and Gui, Shurui and Edwards, Carl and Gao, Nicholas and Ladera, Adriana and Wu, Tailin and Hofgard, Elyssa F. and Tehrani, Aria Mansouri and Wang, Rui and Daigavane, Ameya and Bohde, Montgomery and Kurtin, Jerry and Huang, Qian and Phung, Tuong and Xu, Minkai and Joshi, Chaitanya K. and Mathis, Simon V. and Azizzadenesheli, Kamyar and Fang, Ada and Aspuru-Guzik, Alán and Bekkers, Erik and Bronstein, Michael and Zitnik, Marinka and Anandkumar, Anima and Ermon, Stefano and Liò, Pietro and Yu, Rose and Günnemann, Stephan and Leskovec, Jure and Ji, Heng and Sun, Jimeng and Barzilay, Regina and Jaakkola, Tommi and Coley, Connor W. and Qian, Xiaoning and Qian, Xiaofeng and Smidt, Tess and Ji, Shuiwang},
date = {2023-07-17},
eprint = {2307.08423},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2307.08423},
url = {http://arxiv.org/abs/2307.08423},
urldate = {2023-07-24},
abstract = {Advances in artificial intelligence (AI) are fueling a new paradigm of discoveries in natural sciences. Today, AI has started to advance natural sciences by improving, accelerating, and enabling our understanding of natural phenomena at a wide range of spatial and temporal scales, giving rise to a new area of research known as AI for science (AI4Science). Being an emerging research paradigm, AI4Science is unique in that it is an enormous and highly interdisciplinary area. Thus, a unified and technical treatment of this field is needed yet challenging. This paper aims to provide a technically thorough account of a subarea of AI4Science; namely, AI for quantum, atomistic, and continuum systems. These areas aim at understanding the physical world from the subatomic (wavefunctions and electron density), atomic (molecules, proteins, materials, and interactions), to macro (fluids, climate, and subsurface) scales and form an important subarea of AI4Science. A unique advantage of focusing on these areas is that they largely share a common set of challenges, thereby allowing a unified and foundational treatment. A key common challenge is how to capture physics first principles, especially symmetries, in natural systems by deep learning methods. We provide an in-depth yet intuitive account of techniques to achieve equivariance to symmetry transformations. We also discuss other common technical challenges, including explainability, out-of-distribution generalization, knowledge transfer with foundation and large language models, and uncertainty quantification. To facilitate learning and education, we provide categorized lists of resources that we found to be useful. We strive to be thorough and unified and hope this initial effort may trigger more community interests and efforts to further advance AI4Science.},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,AI4Science,ALIGNN,Allegro,AlphaFold,AML,benchmarking,body-order,CCSD(T),CGCNN,chemistry,Database,DeepH,DFT,DimeNet,drug discovery,E(3),education,EGNN,equivariant,FermiNet,foundation models,G-SchNet,GemNet,generative models,GNN,graph ML,invariance,learning material,library,lists,LLM,M3GNet,MACE,magnetism,MatBench,materials discovery,materials project,MD,MD17,MEGNet,Microsoft Research,ML,ML-DFA,ML-DFT,ML-ESM,ML-FF,ML-QMBP,MLP,model comparison,model taxonomy,molecules,MPNN,NequIP,NQS,OC20,OF-DFT,open questions,out-of-distribution,PAiNN,PauliNet,PDE,PhiSNet,phonon,physics,QM7,QM9,representation learning,resources list,review,review-of-AI4science,review-of-AML,review-of-ML-DFT,roadmap,SchNet,SchNOrb,SE(3),SOTA,SphereNet,spin-dependent,SSL,symmetry,uncertainty quantification,with-code,XAI},
+ pubstate = {prepublished},
+ keywords = {ACE,AI4Science,ALIGNN,Allegro,AlphaFold,AML,benchmarking,body-order,CCSD(T),CGCNN,chemistry,Database,DeepH,DFT,DimeNet,drug discovery,E(3),education,EGNN,equivariant,FermiNet,foundation models,G-SchNet,GemNet,generative models,GNN,graph ML,invariance,learning material,library,lists,LLM,M3GNet,MACE,magnetism,MatBench,materials discovery,materials project,MD,MD17,MEGNet,Microsoft Research,ML,ML-DFA,ML-DFT,ML-ESM,ML-FF,ML-QMBP,MLP,model comparison,model taxonomy,molecules,MPNN,NequIP,NQS,OC20,OF-DFT,open questions,out-of-distribution,PAiNN,PauliNet,PDE,PhiSNet,phonon,physics,QM7,QM9,representation learning,resources list,review,review-of-AI4science,review-of-AML,review-of-ML-DFT,roadmap,SchNet,SchNOrb,SE(3),SOTA,SphereNet,spin-dependent,SSL,symmetry,uncertainty quantification,with-code,XAI},
file = {/Users/wasmer/Nextcloud/Zotero/Zhang et al_2023_Artificial Intelligence for Science in Quantum, Atomistic, and Continuum Systems.pdf;/Users/wasmer/Zotero/storage/PSVYKZKY/2307.html}
}
-@article{zhangCrossoverThreeDimensionalTopological2010,
- title = {Crossover of {{Three-Dimensional Topological Insulator}} of {{Bi2Se3}} to the {{Two-Dimensional Limit}}},
- author = {Zhang, Yi and He, Ke and Chang, Cui-Zu and Song, Can-Li and Wang, Lili and Chen, Xi and Jia, Jinfeng and Fang, Zhong and Dai, Xi and Shan, Wen-Yu and Shen, Shun-Qing and Niu, Qian and Qi, Xiaoliang and Zhang, Shou-Cheng and Ma, Xucun and Xue, Qi-Kun},
- date = {2010-08},
- journaltitle = {Nature Physics},
- shortjournal = {Nature Phys},
- volume = {6},
- number = {8},
- eprint = {0911.3706},
- eprinttype = {arxiv},
- eprintclass = {cond-mat},
- pages = {584--588},
- issn = {1745-2473, 1745-2481},
- doi = {10.1038/nphys1689},
- url = {http://arxiv.org/abs/0911.3706},
- urldate = {2023-07-04},
- abstract = {Bi2Se3 is theoretically predicted1 2and experimentally observed2,3 to be a three dimensional topological insulator. For possible applications, it is important to understand the electronic structure of the planar device. In this work, thickness dependent band structure of molecular beam epitaxy grown ultrathin films of Bi2Se3 is investigated by in situ angle-resolved photoemission spectroscopy. An energy gap is observed for the first time in the topologically protected metallic surface states of bulk Bi2Se3 below the thickness of six quintuple layers, due to the coupling between the surface states from two opposite surfaces of the Bi2Se3 film. The gapped surface states exhibit sizable Rashba-type spin-orbit splitting, due to breaking of structural inversion symmetry induced by SiC substrate. The spin-splitting can be controlled by tuning the potential difference between the two surfaces.},
- keywords = {/unread,Condensed Matter - Materials Science,Condensed Matter - Mesoscale and Nanoscale Physics},
- file = {/Users/wasmer/Nextcloud/Zotero/Zhang et al_2010_Crossover of Three-Dimensional Topological Insulator of Bi2Se3 to the.pdf;/Users/wasmer/Zotero/storage/RH2YKR7S/0911.html}
+@online{zhangArtificialIntelligenceScience2023a,
+ title = {Artificial {{Intelligence}} for {{Science}} in {{Quantum}}, {{Atomistic}}, and {{Continuum Systems}}},
+ author = {Zhang, Xuan and Wang, Limei and Helwig, Jacob and Luo, Youzhi and Fu, Cong and Xie, Yaochen and Liu, Meng and Lin, Yuchao and Xu, Zhao and Yan, Keqiang and Adams, Keir and Weiler, Maurice and Li, Xiner and Fu, Tianfan and Wang, Yucheng and Yu, Haiyang and Xie, YuQing and Fu, Xiang and Strasser, Alex and Xu, Shenglong and Liu, Yi and Du, Yuanqi and Saxton, Alexandra and Ling, Hongyi and Lawrence, Hannah and Stärk, Hannes and Gui, Shurui and Edwards, Carl and Gao, Nicholas and Ladera, Adriana and Wu, Tailin and Hofgard, Elyssa F. and Tehrani, Aria Mansouri and Wang, Rui and Daigavane, Ameya and Bohde, Montgomery and Kurtin, Jerry and Huang, Qian and Phung, Tuong and Xu, Minkai and Joshi, Chaitanya K. and Mathis, Simon V. and Azizzadenesheli, Kamyar and Fang, Ada and Aspuru-Guzik, Alán and Bekkers, Erik and Bronstein, Michael and Zitnik, Marinka and Anandkumar, Anima and Ermon, Stefano and Liò, Pietro and Yu, Rose and Günnemann, Stephan and Leskovec, Jure and Ji, Heng and Sun, Jimeng and Barzilay, Regina and Jaakkola, Tommi and Coley, Connor W. and Qian, Xiaoning and Qian, Xiaofeng and Smidt, Tess and Ji, Shuiwang},
+ date = {2023-11-15},
+ eprint = {2307.08423},
+ eprinttype = {arXiv},
+ eprintclass = {physics},
+ doi = {10.48550/arXiv.2307.08423},
+ url = {http://arxiv.org/abs/2307.08423},
+ urldate = {2024-05-29},
+ abstract = {Advances in artificial intelligence (AI) are fueling a new paradigm of discoveries in natural sciences. Today, AI has started to advance natural sciences by improving, accelerating, and enabling our understanding of natural phenomena at a wide range of spatial and temporal scales, giving rise to a new area of research known as AI for science (AI4Science). Being an emerging research paradigm, AI4Science is unique in that it is an enormous and highly interdisciplinary area. Thus, a unified and technical treatment of this field is needed yet challenging. This work aims to provide a technically thorough account of a subarea of AI4Science; namely, AI for quantum, atomistic, and continuum systems. These areas aim at understanding the physical world from the subatomic (wavefunctions and electron density), atomic (molecules, proteins, materials, and interactions), to macro (fluids, climate, and subsurface) scales and form an important subarea of AI4Science. A unique advantage of focusing on these areas is that they largely share a common set of challenges, thereby allowing a unified and foundational treatment. A key common challenge is how to capture physics first principles, especially symmetries, in natural systems by deep learning methods. We provide an in-depth yet intuitive account of techniques to achieve equivariance to symmetry transformations. We also discuss other common technical challenges, including explainability, out-of-distribution generalization, knowledge transfer with foundation and large language models, and uncertainty quantification. To facilitate learning and education, we provide categorized lists of resources that we found to be useful. We strive to be thorough and unified and hope this initial effort may trigger more community interests and efforts to further advance AI4Science.},
+ pubstate = {prepublished},
+ version = {2},
+ keywords = {/unread,ACE,AI4Science,ALIGNN,Allegro,AlphaFold,AML,benchmarking,body-order,CCSD(T),CGCNN,chemistry,Computer Science - Machine Learning,Database,DeepH,DFT,DimeNet,drug discovery,E(3),education,EGNN,equivariant,FermiNet,foundation models,G-SchNet,GemNet,generative models,GNN,graph ML,invariance,learning material,library,lists,LLM,M3GNet,MACE,magnetism,MatBench,materials discovery,materials project,MD,MD17,MEGNet,Microsoft Research,ML,ML-DFA,ML-DFT,ML-ESM,ML-FF,ML-QMBP,MLP,model comparison,model taxonomy,molecules,MPNN,NequIP,NQS,OC20,OF-DFT,open questions,out-of-distribution,PAiNN,PauliNet,PDE,PhiSNet,phonon,physics,Physics - Computational Physics,QM7,QM9,representation learning,resources list,review,review-of-AI4science,review-of-AML,review-of-ML-DFT,roadmap,SchNet,SchNOrb,SE(3),SOTA,SphereNet,spin-dependent,SSL,symmetry,uncertainty quantification,with-code,XAI},
+ file = {/Users/wasmer/Nextcloud/Zotero/Zhang et al_2023_Artificial Intelligence for Science in Quantum, Atomistic, and Continuum Systems2.pdf;/Users/wasmer/Zotero/storage/J2HWXJKJ/2307.html}
}
@article{zhangCrossoverThreedimensionalTopological2010,
@@ -18301,7 +19992,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
abstract = {The gapless surface states of topological insulators could enable quantitatively different types of electronic device. A study of the topological insulating Bi2Se3 thin films finds that a gap in these states opens up in films below a certain thickness. This in turn suggests that in thicker films, gapless states exist on both upper and lower surfaces.},
issue = {8},
langid = {english},
- keywords = {Atomic,Classical and Continuum Physics,Complex Systems,Condensed Matter Physics,general,Mathematical and Computational Physics,Molecular,Optical and Plasma Physics,Physics,Theoretical},
+ keywords = {2D material,ARPES,Bi2Se3,condensed matter,mesoscopic,physics,SOC,thin film,topological,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Zhang et al_2010_Crossover of the three-dimensional topological insulator Bi2Se3 to the.pdf}
}
@@ -18342,13 +20033,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhang, Duo and Liu, Xinzijian and Zhang, Xiangyu and Zhang, Chengqian and Cai, Chun and Bi, Hangrui and Du, Yiming and Qin, Xuejian and Huang, Jiameng and Li, Bowen and Shan, Yifan and Zeng, Jinzhe and Zhang, Yuzhi and Liu, Siyuan and Li, Yifan and Chang, Junhan and Wang, Xinyan and Zhou, Shuo and Liu, Jianchuan and Luo, Xiaoshan and Wang, Zhenyu and Jiang, Wanrun and Wu, Jing and Yang, Yudi and Yang, Jiyuan and Yang, Manyi and Gong, Fu-Qiang and Zhang, Linshuang and Shi, Mengchao and Dai, Fu-Zhi and York, Darrin M. and Liu, Shi and Zhu, Tong and Zhong, Zhicheng and Lv, Jian and Cheng, Jun and Jia, Weile and Chen, Mohan and Ke, Guolin and E, Weinan and Zhang, Linfeng and Wang, Han},
date = {2023-12-24},
eprint = {2312.15492},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2312.15492},
url = {http://arxiv.org/abs/2312.15492},
urldate = {2024-05-06},
abstract = {The rapid development of artificial intelligence (AI) is driving significant changes in the field of atomic modeling, simulation, and design. AI-based potential energy models have been successfully used to perform large-scale and long-time simulations with the accuracy of ab initio electronic structure methods. However, the model generation process still hinders applications at scale. We envision that the next stage would be a model-centric ecosystem, in which a large atomic model (LAM), pre-trained with as many atomic datasets as possible and can be efficiently fine-tuned and distilled to downstream tasks, would serve the new infrastructure of the field of molecular modeling. We propose DPA-2, a novel architecture for a LAM, and develop a comprehensive pipeline for model fine-tuning, distillation, and application, associated with automatic workflows. We show that DPA-2 can accurately represent a diverse range of chemical systems and materials, enabling high-quality simulations and predictions with significantly reduced efforts compared to traditional methods. Our approach paves the way for a universal large atomic model that can be widely applied in molecular and material simulation research, opening new opportunities for scientific discoveries and industrial applications.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ablation study,Allegro,AML,attention,benchmarking,dataset,DeePMD-kit,Equiformer,equivariant,fine-tuning,foundation models,GemNet,generalization,ML,MLP,multi-task learning,NequIP,periodic table,pretrained models,property prediction,transfer learning,transformer,universal potential,with-code,with-data,zero-shot generalization},
file = {/Users/wasmer/Nextcloud/Zotero/Zhang et al_2023_DPA-2.pdf;/Users/wasmer/Zotero/storage/LAGXG4QQ/2312.html}
}
@@ -18427,19 +20118,38 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
file = {/Users/wasmer/Nextcloud/Zotero/Zhang et al_2018_Machine Learning Topological Invariants with Neural Networks.pdf;/Users/wasmer/Zotero/storage/XCPMLTVF/PhysRevLett.120.html}
}
+@article{zhangMagneticTopologicalMaterials2023,
+ title = {Magnetic Topological Materials in Two-Dimensional: Theory, Material Realization and Application Prospects},
+ shorttitle = {Magnetic Topological Materials in Two-Dimensional},
+ author = {Zhang, Xiaoming and Wang, Xiaotian and He, Tingli and Wang, Lirong and Yu, Wei-Wang and Liu, Ying and Liu, Guodong and Cheng, Zhenxiang},
+ date = {2023-11-15},
+ journaltitle = {Science Bulletin},
+ shortjournal = {Science Bulletin},
+ volume = {68},
+ number = {21},
+ pages = {2639--2657},
+ issn = {2095-9273},
+ doi = {10.1016/j.scib.2023.09.004},
+ url = {https://www.sciencedirect.com/science/article/pii/S2095927323006205},
+ urldate = {2024-05-19},
+ abstract = {Two-dimensional (2D) magnetism and nontrivial band topology are both areas of research that are currently receiving significant attention in the study of 2D materials. Recently, a novel class of materials has emerged, known as 2D magnetic topological materials, which elegantly combine 2D magnetism and nontrivial topology. This field has garnered increasing interest, especially due to the emergence of several novel magnetic topological states that have been generalized into the 2D scale. These states include antiferromagnetic topological insulators/semimetals, second-order topological insulators, and topological half-metals. Despite the rapid advancements in this emerging research field in recent years, there have been few comprehensive summaries of the state-of-the-art progress. Therefore, this review aims to provide a thorough analysis of current progress on 2D magnetic topological materials. We cover various 2D magnetic topological insulators, a range of 2D magnetic topological semimetals, and the novel 2D topological half-metals, systematically analyzing the basic topological theory, the course of development, the material realization, and potential applications. Finally, we discuss the challenges and prospects for 2D magnetic topological materials, highlighting the potential for future breakthroughs in this exciting field.},
+ keywords = {/unread,2D material,condensed matter,good figures,half-metals,Hall effect,Hall QAHE,physics,quantum materials,review,review-of-TIs,semimetal,spintronics,superconductor,topological,topological insulator,Topological Superconductor},
+ file = {/Users/wasmer/Zotero/storage/3CMDQH6K/S2095927323006205.html}
+}
+
@online{zhangPushingLimitsAtomistic2019,
title = {Pushing the Limits of Atomistic Simulations towards Ultra-High Temperature: A Machine-Learning Force Field for {{ZrB2}}},
shorttitle = {Pushing the Limits of Atomistic Simulations towards Ultra-High Temperature},
author = {Zhang, Yanhui and Lunghi, Alessandro and Sanvito, Stefano},
date = {2019-11-08},
eprint = {1911.03307},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:quant-ph},
doi = {10.48550/arXiv.1911.03307},
url = {http://arxiv.org/abs/1911.03307},
urldate = {2023-02-23},
abstract = {Determining thermal and physical quantities across a broad temperature domain, especially up to the ultra-high temperature region, is a formidable theoretical and experimental challenge. At the same time it is essential for understanding the performance of ultra-high temperature ceramic (UHTC) materials. Here we present the development of a machine-learning force field for ZrB2, one of the primary members of the UHTC family with a complex bonding structure. The force field exhibits chemistry accuracy for both energies and forces and can reproduce structural, elastic and phonon properties, including thermal expansion and thermal transport. A thorough comparison with available empirical potentials shows that our force field outperforms the competitors. Most importantly, its effectiveness is extended from room temperature to the ultra-high temperature region (up to \textasciitilde{} 2,500 K), where measurements are very difficult, costly and some time impossible. Our work demonstrates that machine-learning force fields can be used for simulations of materials in a harsh environment, where no experimental tools are available, but crucial for a number of engineering applications, such as in aerospace, aviation and nuclear.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,\_tablet,Condensed Matter - Materials Science,Quantum Physics},
file = {/Users/wasmer/Nextcloud/Zotero/Zhang et al_2019_Pushing the limits of atomistic simulations towards ultra-high temperature.pdf;/Users/wasmer/Zotero/storage/IIZUJ8Y9/1911.html}
}
@@ -18513,7 +20223,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
urldate = {2023-06-23},
abstract = {Topological insulators are new quantum states with helical gapless edge or surface states inside the bulk band gap. These topological surface states are robust against weak time-reversal invariant perturbations without closing the bulk band gap, such as lattice distortions and non-magnetic impurities. Recently a variety of topological insulators have been predicted by theories, and observed by experiments. First-principles calculations have been widely used to predict topological insulators with great success. In this review, we summarize the current progress in this field from the perspective of first-principles calculations. First of all, the basic concepts of topological insulators and the frequently-used techniques within first-principles calculations are briefly introduced. Secondly, we summarize general methodologies to search for new topological insulators. In the last part, based on the band inversion picture first introduced in the context of HgTe, we classify topological insulators into three types with s–p, p–p and d–f, and discuss some representative examples for each type. Surface states of topological insulator Bi2Se3 consist of a single Dirac cone, as obtained from first-principles calculations. (© 2013 WILEY-VCH Verlag GmbH \& Co. KGaA, Weinheim)},
langid = {english},
- keywords = {\_tablet,DFT,DMFT,Hall effect,Hall QAHE,LDA,LDA+DMFT,LDA+U,review,SOC,topological insulator},
+ keywords = {DFT,DMFT,Hall effect,Hall QAHE,LDA,LDA+DMFT,LDA+U,review,SOC,topological insulator},
file = {/Users/wasmer/Nextcloud/Zotero/Zhang_Zhang_2013_Topological insulators from the perspective of first-principles calculations.pdf;/Users/wasmer/Zotero/storage/ZU36AAB4/pssr.html}
}
@@ -18571,14 +20281,14 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhong, Yang and Zhang, Binhua and Yu, Hongyu and Gong, Xingao and Xiang, Hongjun},
date = {2023-06-02},
eprint = {2306.01558},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2306.01558},
url = {http://arxiv.org/abs/2306.01558},
urldate = {2023-06-12},
abstract = {Complex spin-spin interactions in magnets can often lead to magnetic superlattices with complex local magnetic arrangements, and many of the magnetic superlattices have been found to possess non-trivial topological electronic properties. Due to the huge size and complex magnetic moment arrangement of the magnetic superlattices, it is a great challenge to perform a direct DFT calculation on them. In this work, an equivariant deep learning framework is designed to accelerate the electronic calculation of magnetic systems by exploiting both the equivariant constraints of the magnetic Hamiltonian matrix and the physical rules of spin-spin interactions. This framework can bypass the costly self-consistent iterations and build a direct mapping from a magnetic configuration to the ab initio Hamiltonian matrix. After training on the magnets with random magnetic configurations, our model achieved high accuracy on the test structures outside the training set, such as spin spiral and non-collinear antiferromagnetic configurations. The trained model is also used to predict the energy bands of a skyrmion configuration of NiBrI containing thousands of atoms, showing the high efficiency of our model on large magnetic superlattices.},
- pubstate = {preprint},
- keywords = {\_tablet,2D material,AFM,AML,DFT,DFT speedup,DFT speedup with ML,Dzyaloshinskii–Moriya interaction,E(3),equivariant,GNN,Hall effect,Heisenberg model,higher-order exchange interactions,iron,Jij,magnetic Hamiltonian,magnetic supperlattice,magnetism,ML,ML-DFT,ML-ESM,MPNN,non-collinear,OpenMX,prediction from magnetic configuration,prediction of Hamiltonian matrix,prediction of Jij,skyrmions,SO(3),SOC,spin spiral,spin-dependent,SU(2),ternary systems,TRS},
+ pubstate = {prepublished},
+ keywords = {2D material,AFM,AML,DFT,DFT speedup,DFT speedup with ML,Dzyaloshinskii–Moriya interaction,E(3),equivariant,GNN,Hall effect,Heisenberg model,higher-order exchange interactions,iron,Jij,magnetic Hamiltonian,magnetic supperlattice,magnetism,ML,ML-DFT,ML-ESM,MPNN,non-collinear,OpenMX,prediction from magnetic configuration,prediction of Hamiltonian matrix,prediction of Jij,skyrmions,SO(3),SOC,spin spiral,spin-dependent,SU(2),ternary systems,TRS},
file = {/Users/wasmer/Nextcloud/Zotero/Zhong et al_2023_Accelerating the electronic-structure calculation of magnetic systems by.pdf;/Users/wasmer/Zotero/storage/RJIQYZHY/2306.html}
}
@@ -18587,13 +20297,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhong, Yang and Yu, Hongyu and Gong, Xingao and Xiang, Hongjun},
date = {2022-01-15},
eprint = {2201.05770},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2201.05770},
url = {http://arxiv.org/abs/2201.05770},
urldate = {2024-04-18},
abstract = {Message-passing neural networks (MPNN) have shown extremely high efficiency and accuracy in predicting the physical properties of molecules and crystals, and are expected to become the next-generation material simulation tool after the density functional theory (DFT). However, there is currently a lack of a general MPNN framework for directly predicting the tensor properties of the crystals. In this work, a general framework for the prediction of tensor properties was proposed: the tensor property of a crystal can be decomposed into the average of the tensor contributions of all the atoms in the crystal, and the tensor contribution of each atom can be expanded as the sum of the tensor projections in the directions of the edges connecting the atoms. On this basis, the edge-based expansions of force vectors, Born effective charges (BECs), dielectric (DL) and piezoelectric (PZ) tensors were proposed. These expansions are rotationally equivariant, while the coefficients in these tensor expansions are rotationally invariant scalars which are similar to physical quantities such as formation energy and band gap. The advantage of this tensor prediction framework is that it does not require the network itself to be equivariant. Therefore, in this work, we directly designed the edge-based tensor prediction graph neural network (ETGNN) model on the basis of the invariant graph neural network to predict tensors. The validity and high precision of this tensor prediction framework were shown by the tests of ETGNN on the extended systems, random perturbed structures and JARVIS-DFT datasets. This tensor prediction framework is general for nearly all the GNNs and can achieve higher accuracy with more advanced GNNs in the future.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,equivariant,GNN,HEA,invariance,JARVIS-DFT,materials,ML,MPNN,prediction of charge,prediction of forces,tensorial target},
file = {/Users/wasmer/Nextcloud/Zotero/Zhong et al_2022_Edge-based Tensor prediction via graph neural networks.pdf;/Users/wasmer/Zotero/storage/RQ2TNRCK/2201.html}
}
@@ -18610,7 +20320,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
url = {https://doi.org/10.1021/acs.jpclett.3c01200},
urldate = {2023-07-13},
abstract = {Graph neural networks (GNNs) have been shown to be extremely flexible and accurate in predicting the physical properties of molecules and crystals. However, traditional invariant GNNs are not compatible with directional properties, which currently limits their usage to the prediction of only invariant scalar properties. To address this issue, here we propose a general framework, i.e., an edge-based tensor prediction graph neural network, in which a tensor is expressed as the linear combination of the local spatial components projected on the edge directions of clusters with varying sizes. This tensor decomposition is rotationally equivariant and exactly satisfies the symmetry of the local structures. The accuracy and universality of our new framework are demonstrated by the successful prediction of various tensor properties from first to third order. The framework proposed in this work will enable GNNs to step into the broad field of prediction of directional properties.},
- keywords = {\_tablet,AML,benchmarking,DeePMD-kit,DimeNet++,equivariant,FieldSchNet,GNN,GPR,JARVIS-DFT,magnetic anisotropy,magnetism,ML,MLP,MPNN,PCA,prediction of magnetic anisotropy,SA-GPR,SchNetPack,SOAP,tensorial target},
+ keywords = {AML,benchmarking,DeePMD-kit,DimeNet++,equivariant,FieldSchNet,GNN,GPR,JARVIS-DFT,magnetic anisotropy,magnetism,ML,MLP,MPNN,PCA,prediction of magnetic anisotropy,SA-GPR,SchNetPack,SOAP,tensorial target},
file = {/Users/wasmer/Nextcloud/Zotero/false;/Users/wasmer/Nextcloud/Zotero/Zhong et al_2023_A General Tensor Prediction Framework Based on Graph Neural Networks.pdf;/Users/wasmer/Zotero/storage/B7FX9ZP8/acs.jpclett.html}
}
@@ -18631,7 +20341,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
abstract = {This work presents an E(3) equivariant graph neural network called HamGNN, which can fit the electronic Hamiltonian matrix of molecules and solids by a complete data-driven method. Unlike invariant models that achieve equivariance approximately through data augmentation, HamGNN employs E(3) equivariant convolutions to construct the Hamiltonian matrix, ensuring strict adherence to all equivariant constraints inherent in the physical system. In contrast to previous models with limited transferability, HamGNN demonstrates exceptional accuracy on various datasets, including QM9 molecular datasets, carbon allotropes, silicon allotropes, SiO2 isomers, and BixSey compounds. The trained HamGNN models exhibit accurate predictions of electronic structures for large crystals beyond the training set, including the Moiré twisted bilayer MoS2 and silicon supercells with dislocation defects, showcasing remarkable transferability and generalization capabilities. The HamGNN model, trained on small systems, can serve as an efficient alternative to density functional theory (DFT) for accurately computing the electronic structures of large systems.},
issue = {1},
langid = {english},
- keywords = {\_tablet,AML,bismuth selenide,DFT,e3nn,GNN,HamGNN,library,line defects,ML,ML-DFT,ML-ESM,MoS2,MPNN,OpenMX,prediction of Hamiltonian matrix,PyTorch,SOC,spin-dependent,SU(2),TB,tight binding,TMDC,TRS,twisted bilayer,VASP,with-code,with-data},
+ keywords = {AML,bismuth selenide,DFT,e3nn,GNN,HamGNN,library,line defects,ML,ML-DFT,ML-ESM,MoS2,MPNN,OpenMX,prediction of Hamiltonian matrix,PyTorch,SOC,spin-dependent,SU(2),TB,tight binding,TMDC,TRS,twisted bilayer,VASP,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Zhong et al_2023_Transferable equivariant graph neural networks for the Hamiltonians of.pdf}
}
@@ -18640,13 +20350,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhong, Yang and Yu, Hongyu and Su, Mao and Gong, Xingao and Xiang, Hongjun},
date = {2023-02-04},
eprint = {2210.16190},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.1038/s41524-023-01130-4},
url = {http://arxiv.org/abs/2210.16190},
urldate = {2024-04-18},
abstract = {Using the message-passing mechanism in machine learning (ML) instead of self-consistent iterations to directly build the mapping from structures to electronic Hamiltonian matrices will greatly improve the efficiency of density functional theory (DFT) calculations. In this work, we proposed a general analytic Hamiltonian representation in an E(3) equivariant framework, which can fit the ab initio Hamiltonian of molecules and solids by a complete data-driven method and are equivariant under rotation, space inversion, and time reversal operations. Our model reached state-of-the-art precision in the benchmark test and accurately predicted the electronic Hamiltonian matrices and related properties of various periodic and aperiodic systems, showing high transferability and generalization ability. This framework provides a general transferable model that can be used to accelerate the electronic structure calculations on different large systems with the same network weights trained on small structures.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {AML,bismuth selenide,DFT,e3nn,GNN,HamGNN,library,ML,ML-DFT,ML-ESM,MoS2,MPNN,OpenMX,prediction of Hamiltonian matrix,PyTorch,todo-tagging,twisted bilayer,VASP,with-code,with-data},
file = {/Users/wasmer/Nextcloud/Zotero/Zhong et al_2023_Transferable E(3) equivariant parameterization for Hamiltonian of molecules and2.pdf;/Users/wasmer/Zotero/storage/SCZR9M92/2210.html}
}
@@ -18656,14 +20366,14 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhong, Yang and Yu, Hongyu and Yang, Jihui and Guo, Xingyu and Xiang, Hongjun and Gong, Xingao},
date = {2024-04-15},
eprint = {2402.09251},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cond-mat, physics:physics},
doi = {10.48550/arXiv.2402.09251},
url = {http://arxiv.org/abs/2402.09251},
urldate = {2024-04-18},
abstract = {While density functional theory (DFT) serves as a prevalent computational approach in electronic structure calculations, its computational demands and scalability limitations persist. Recently, leveraging neural networks to parameterize the Kohn-Sham DFT Hamiltonian has emerged as a promising avenue for accelerating electronic structure computations. Despite advancements, challenges such as the necessity for computing extensive DFT training data to explore each new system and the complexity of establishing accurate ML models for multi-elemental materials still exist. Addressing these hurdles, this study introduces a universal electronic Hamiltonian model trained on Hamiltonian matrices obtained from first-principles DFT calculations of nearly all crystal structures on the Materials Project. We demonstrate its generality in predicting electronic structures across the whole periodic table, including complex multi-elemental systems, solid-state electrolytes, Moir\textbackslash 'e twisted bilayer heterostructure, and metal-organic frameworks (MOFs). Moreover, we utilize the universal model to conduct high-throughput calculations of electronic structures for crystals in GeNOME datasets, identifying 3,940 crystals with direct band gaps and 5,109 crystals with flat bands. By offering a reliable efficient framework for computing electronic properties, this universal Hamiltonian model lays the groundwork for advancements in diverse fields, such as easily providing a huge data set of electronic structures and also making the materials design across the whole periodic table possible.},
- pubstate = {preprint},
- keywords = {\_tablet,ACE,AML,chemical species scaling problem,DeepH,DFT,foundation models,GNoME,HamGNN,HEA,heterostructures,materials,materials project,ML,ML-DFT,ML-ESM,OpenMX,PhiSNet,prediction from energy,prediction from structure,prediction of Hamiltonian matrix,SchNOrb,twisted bilayer,with-code},
+ pubstate = {prepublished},
+ keywords = {ACE,AML,chemical species scaling problem,DeepH,DFT,foundation models,GNoME,HamGNN,HEA,heterostructures,materials,materials project,ML,ML-DFT,ML-ESM,OpenMX,PhiSNet,prediction from energy,prediction from structure,prediction of Hamiltonian matrix,SchNOrb,twisted bilayer,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Zhong et al_2024_Universal Machine Learning Kohn-Sham Hamiltonian for Materials.pdf;/Users/wasmer/Zotero/storage/QIP9JFLF/2402.html}
}
@@ -18673,13 +20383,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhou, Ce and Li, Qian and Li, Chen and Yu, Jun and Liu, Yixin and Wang, Guangjing and Zhang, Kai and Ji, Cheng and Yan, Qiben and He, Lifang and Peng, Hao and Li, Jianxin and Wu, Jia and Liu, Ziwei and Xie, Pengtao and Xiong, Caiming and Pei, Jian and Yu, Philip S. and Sun, Lichao},
date = {2023-03-30},
eprint = {2302.09419},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs},
doi = {10.48550/arXiv.2302.09419},
url = {http://arxiv.org/abs/2302.09419},
urldate = {2023-04-14},
abstract = {Pretrained Foundation Models (PFMs) are regarded as the foundation for various downstream tasks with different data modalities. A PFM (e.g., BERT, ChatGPT, and GPT-4) is trained on large-scale data which provides a reasonable parameter initialization for a wide range of downstream applications. BERT learns bidirectional encoder representations from Transformers, which are trained on large datasets as contextual language models. Similarly, the generative pretrained transformer (GPT) method employs Transformers as the feature extractor and is trained using an autoregressive paradigm on large datasets. Recently, ChatGPT shows promising success on large language models, which applies an autoregressive language model with zero shot or few shot prompting. The remarkable achievements of PFM have brought significant breakthroughs to various fields of AI. Numerous studies have proposed different methods, raising the demand for an updated survey. This study provides a comprehensive review of recent research advancements, challenges, and opportunities for PFMs in text, image, graph, as well as other data modalities. The review covers the basic components and existing pretraining methods used in natural language processing, computer vision, and graph learning. Additionally, it explores advanced PFMs used for different data modalities and unified PFMs that consider data quality and quantity. The review also discusses research related to the fundamentals of PFMs, such as model efficiency and compression, security, and privacy. Finally, the study provides key implications, future research directions, challenges, and open problems in the field of PFMs. Overall, this survey aims to shed light on the research of the PFMs on scalability, security, logical reasoning ability, cross-domain learning ability, and the user-friendly interactive ability for artificial general intelligence.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,BERT,ChatGPT,few-shot learning,foundation models,GATN,GCN,General ML,GNN,GPT,graph ML,LLM,ML,transfer learning,transformer,zero-shot learning},
file = {/Users/wasmer/Nextcloud/Zotero/Zhou et al_2023_A Comprehensive Survey on Pretrained Foundation Models.pdf;/Users/wasmer/Zotero/storage/CWZ9H6CB/2302.html}
}
@@ -18689,13 +20399,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhou, Yi and Barnes, Connelly and Lu, Jingwan and Yang, Jimei and Li, Hao},
date = {2020-06-08},
eprint = {1812.07035},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {cs, stat},
doi = {10.48550/arXiv.1812.07035},
url = {http://arxiv.org/abs/1812.07035},
urldate = {2024-04-05},
abstract = {In neural networks, it is often desirable to work with various representations of the same space. For example, 3D rotations can be represented with quaternions or Euler angles. In this paper, we advance a definition of a continuous representation, which can be helpful for training deep neural networks. We relate this to topological concepts such as homeomorphism and embedding. We then investigate what are continuous and discontinuous representations for 2D, 3D, and n-dimensional rotations. We demonstrate that for 3D rotations, all representations are discontinuous in the real Euclidean spaces of four or fewer dimensions. Thus, widely used representations such as quaternions and Euler angles are discontinuous and difficult for neural networks to learn. We show that the 3D rotations have continuous representations in 5D and 6D, which are more suitable for learning. We also present continuous representations for the general case of the n-dimensional rotation group SO(n). While our main focus is on rotations, we also show that our constructions apply to other groups such as the orthogonal group and similarity transforms. We finally present empirical results, which show that our continuous rotation representations outperform discontinuous ones for several practical problems in graphics and vision, including a simple autoencoder sanity test, a rotation estimator for 3D point clouds, and an inverse kinematics solver for 3D human poses.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {/unread,general ML,geometric deep learning,group theory,ML,rec-by-katsumoto,rotational symmetry,with-code},
file = {/Users/wasmer/Nextcloud/Zotero/Zhou et al_2020_On the Continuity of Rotation Representations in Neural Networks.pdf;/Users/wasmer/Zotero/storage/IXG2BA7I/1812.html}
}
@@ -18731,7 +20441,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
url = {https://www.pnas.org/doi/10.1073/pnas.1801181115},
urldate = {2023-07-12},
abstract = {Exciting advances have been made in artificial intelligence (AI) during recent decades. Among them, applications of machine learning (ML) and deep learning techniques brought human-competitive performances in various tasks of fields, including image recognition, speech recognition, and natural language understanding. Even in Go, the ancient game of profound complexity, the AI player has already beat human world champions convincingly with and without learning from the human. In this work, we show that our unsupervised machines (Atom2Vec) can learn the basic properties of atoms by themselves from the extensive database of known compounds and materials. These learned properties are represented in terms of high-dimensional vectors, and clustering of atoms in vector space classifies them into meaningful groups consistent with human knowledge. We use the atom vectors as basic input units for neural networks and other ML models designed and trained to predict materials properties, which demonstrate significant accuracy.},
- keywords = {/unread,AML,Atom2Vec,compositional descriptors,descriptors,embedding,library,materials,ML,unsupervised learning,with-code,Word2Vec},
+ keywords = {/unread,AML,Atom2Vec,compositional descriptors,descriptors,embedding,language models,library,materials,ML,unsupervised learning,with-code,Word2Vec},
file = {/Users/wasmer/Zotero/storage/HN96KIR2/Zhou et al. - 2018 - Learning atoms for materials discovery.pdf}
}
@@ -18740,13 +20450,13 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
author = {Zhou, Dexuan and Chen, Huajie and Ho, Cheuk Hin and Ortner, Christoph},
date = {2023-05-04},
eprint = {2304.04260},
- eprinttype = {arxiv},
+ eprinttype = {arXiv},
eprintclass = {physics},
doi = {10.48550/arXiv.2304.04260},
url = {http://arxiv.org/abs/2304.04260},
urldate = {2023-12-18},
abstract = {The atomic cluster expansion (ACE) (Drautz, 2019) yields a highly efficient and intepretable parameterisation of symmetric polynomials that has achieved great success in modelling properties of many-particle systems. In the present work we extend the practical applicability of the ACE framework to the computation of many-electron wave functions. To that end, we develop a customized variational Monte-Carlo algorithm that exploits the sparsity and hierarchical properties of ACE wave functions. We demonstrate the feasibility on a range of proof-of-concept applications to one-dimensional systems.},
- pubstate = {preprint},
+ pubstate = {prepublished},
keywords = {ACE,AML,descriptors,ML,ML-ESM,ML-WFT,VMC},
file = {/Users/wasmer/Nextcloud/Zotero/Zhou et al_2023_A Multilevel Method for Many-Electron Schr- o dinger Equations Based on the.pdf;/Users/wasmer/Zotero/storage/9V3IPFCI/2304.html}
}
@@ -18768,6 +20478,42 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
file = {/Users/wasmer/Nextcloud/Zotero/Zhu_2016_Bogoliubov-de Gennes Method and Its Applications.pdf}
}
+@article{zillsCollaborationMachineLearnedPotentials2024,
+ title = {Collaboration on {{Machine-Learned Potentials}} with {{IPSuite}}: {{A Modular Framework}} for {{Learning-on-the-Fly}}},
+ shorttitle = {Collaboration on {{Machine-Learned Potentials}} with {{IPSuite}}},
+ author = {Zills, Fabian and Schäfer, Moritz René and Segreto, Nico and Kästner, Johannes and Holm, Christian and Tovey, Samuel},
+ date = {2024-04-18},
+ journaltitle = {The Journal of Physical Chemistry B},
+ shortjournal = {J. Phys. Chem. B},
+ volume = {128},
+ number = {15},
+ pages = {3662--3676},
+ publisher = {American Chemical Society},
+ issn = {1520-6106},
+ doi = {10.1021/acs.jpcb.3c07187},
+ url = {https://doi.org/10.1021/acs.jpcb.3c07187},
+ urldate = {2024-06-14},
+ abstract = {The field of machine learning potentials has experienced a rapid surge in progress, thanks to advances in machine learning theory, algorithms, and hardware capabilities. While the underlying methods are continuously evolving, the infrastructure for their deployment has lagged. The community, due to these rapid developments, frequently finds itself split into groups built around different implementations of machine-learned potentials. In this work, we introduce IPSuite, a Python-driven software package designed to connect different methods and algorithms from the comprehensive field of machine-learned potentials into a single platform while also providing a collaborative infrastructure, helping ensure reproducibility. Furthermore, the data management infrastructure of the IPSuite code enables simple model sharing and deployment in simulations. Currently, IPSuite supports six state-of-the-art machine learning approaches for the fitting of interatomic potentials as well as a variety of methods for the selection of training data, running of ab initio calculations, learning-on-the-fly strategies, model evaluation, and simulation deployment.},
+ keywords = {/unread,active learning,active learning protocol,Allegro,AML,ANI,ase,benchmarking,database generation,DVC,GAP,Jax-MD,LAMMPS,library,MACE,MD,ML,MLOps,MLP,model deployment,NequIP,RDM,scientific workflows,SMILES,uncertainty quantification,version control,with-code,Zincware},
+ file = {/Users/wasmer/Nextcloud/Zotero/Zills et al_2024_Collaboration on Machine-Learned Potentials with IPSuite.pdf}
+}
+
+@online{zillsZnTrackDataCode2024,
+ title = {{{ZnTrack}} -- {{Data}} as {{Code}}},
+ author = {Zills, Fabian and Schäfer, Moritz and Tovey, Samuel and Kästner, Johannes and Holm, Christian},
+ date = {2024-01-19},
+ eprint = {2401.10603},
+ eprinttype = {arXiv},
+ eprintclass = {cs},
+ doi = {10.48550/arXiv.2401.10603},
+ url = {http://arxiv.org/abs/2401.10603},
+ urldate = {2024-06-14},
+ abstract = {The past decade has seen tremendous breakthroughs in computation and there is no indication that this will slow any time soon. Machine learning, large-scale computing resources, and increased industry focus have resulted in rising investments in computer-driven solutions for data management, simulations, and model generation. However, with this growth in computation has come an even larger expansion of data and with it, complexity in data storage, sharing, and tracking. In this work, we introduce ZnTrack, a Python-driven data versioning tool. ZnTrack builds upon established version control systems to provide a user-friendly and easy-to-use interface for tracking parameters in experiments, designing workflows, and storing and sharing data. From this ability to reduce large datasets to a simple Python script emerges the concept of Data as Code, a core component of the work presented here and an undoubtedly important concept as the age of computation continues to evolve. ZnTrack offers an open-source, FAIR data compatible Python package to enable users to harness these concepts of the future.},
+ pubstate = {prepublished},
+ keywords = {/unread,DVC,FAIR,library,metadata,MLOps,RDM,RSE,scientific workflows,version control,with-code,Zincware},
+ file = {/Users/wasmer/Nextcloud/Zotero/Zills et al_2024_ZnTrack -- Data as Code.pdf;/Users/wasmer/Zotero/storage/IENETMSA/2401.html}
+}
+
@thesis{zimmermannInitioDescriptionTransverse2014,
title = {Ab Initio Description of Transverse Transport Due to Impurity Scattering in Transition-Metals},
author = {Zimmermann, Bernd},
@@ -18779,7 +20525,7 @@ Subject\_term\_id: computational-methods;corrosion;mathematics-and-computing;the
abstract = {This thesis attempts to shed light on various spin-orbit driven transport phenomenain materials, as a crucial for the further development of the field of spintronics. Inparticular, we address the skew-scattering mechanism in dilute alloys, which gives rise to the anomalous and spin Hall effect, as well as spin-relaxation processes. We create the tools to access these quantities from \$\textbackslash textit\{ab initio\}\$ calculations in the framework of the full-potential all-electron Korringa-Kohn-Rostoker Green-function method, by (a) developing and implementing a new tetrahedron method for the calculation of complicated, multi-sheeted Fermi surfaces even of complex transition-metal compounds, and (b) developing an efficiently parallelized and thus highly scalable computer program (up to thousands of processors) for the precise calculation of scattering properties. In a first application of the new tetrahedron method, we calculate the Elliott-Yafet spin-mixing parameter on the Fermi surfaces of 5\$\textbackslash textit\{d\}\$ and 6\$\textbackslash textit\{sp\}\$ metals, and discover a yet unexplored dependence on the electron's spin-polarization direction. As we show, this anisotropy can reach gigantic values in uniaxial hcp crystals due to the emergenceof large spin-ip hot-areas or hot-loops on the Fermi surface, supported by the low symmetry of the hcp crystal. A simple model is able to reveal an interesting interplay between the orbital character of the states at special points, lines or areas in the Brillouin zone and the matrix-elements of the spin-flip part of the spin-orbit coupling operator. We further calculate the skew-scattering contribution to the anomalous Hall effect(AHE) in dilute alloys based on a ferromagnetic host for the first time. A systematic study of 3\$\textbackslash textit\{d\}\$ impurities in bcc Fe, as well as the non-magnetic hosts Pd, Pt and Au, allows us to identify trends across the periodic table. In all our calculations, we also observe a strong correlation between the spin Hall effect and anomalous Hall effect in these materials, which is of interest for the creation and detection of strongly spin-polarized currents. A Fermi-surface analysis of the contributions to the AHE reveals a non-trivial, peaked behavior at small hot-spots around spin-orbit lifted degeneracies. We then proceed to the more complicated \$\textbackslash textit\{L\}\$1\$\_\{0\}\$-ordered alloy FePt and address different kinds of disorder. We showcase the power of our method by treating the very complicated compounds Fe\$\_\{x\}\$Mn\$\_\{1-x\}\$Si and MnSi\$\_\{1-x\}\$Ge\$\_\{x\}\$, based on the non-Fermi liquid manganese silicide (MnSi). Finally, we also calculate the pure spin Hall effect for 4\$\textbackslash textit\{d\}\$/5\$\textbackslash textit\{sp\}\$ and 5\$\textbackslash textit\{d\}\$/6\$\textbackslash textit\{sp\}\$ impurities in fcc Ir and hcp Re hosts. For the latter, we discover a strong dependence on the electron's spin-polarization direction. Zimmermann, Bernd},
isbn = {9783893369850},
langid = {english},
- keywords = {\_tablet,Boltzmann theory,Hall AHE,Hall effect,Hall SHE,juKKR,KKR,PGI-1/IAS-1,thesis,transport properties},
+ keywords = {Boltzmann theory,Hall AHE,Hall effect,Hall SHE,juKKR,KKR,PGI-1/IAS-1,thesis,transport properties},
file = {/Users/wasmer/Nextcloud/Zotero/Zimmermann_2014_Ab initio description of transverse transport due to impurity scattering in.pdf;/Users/wasmer/Zotero/storage/QL7I6VYG/171881.html}
}
--
GitLab