The MLIP package: moment tensor potentials with MPI and active learning

被引：490

作者：

Novikov, Ivan S. ^{[1
]}

Gubaev, Konstantin ^{[1
,2
]}

Podryabinkin, Evgeny, V ^{[1
]}

Shapeev, Alexander, V ^{[1
]}

机构：

[1] Skolkovo Innovat Ctr, Skolkovo Inst Sci & Technol, Nobel St 3, Moscow 143026, Russia

[2] Delft Univ Technol, Dept Mat Sci & Engn, Mekelweg 2, NL-2628 CD Delft, Netherlands

来源：

MACHINE LEARNING-SCIENCE AND TECHNOLOGY | 2021年 / 2卷 / 02期

基金：

俄罗斯科学基金会;

关键词：

machine-learning interatomic potentials; active learning; ab initio calculations; TOTAL-ENERGY CALCULATIONS; INTERATOMIC POTENTIALS; THERMAL-CONDUCTIVITY; SEMICONDUCTORS; ACCURACY; DYNAMICS; ALLOYS;

D O I：

10.1088/2632-2153/abc9fe

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

The subject of this paper is the technology (the 'how') of constructing machine-learning interatomic potentials, rather than science (the 'what' and 'why') of atomistic simulations using machine-learning potentials. Namely, we illustrate how to construct moment tensor potentials using active learning as implemented in the MLIP package, focusing on the efficient ways to automatically sample configurations for the training set, how expanding the training set changes the error of predictions, how to set up ab initio calculations in a cost-effective manner, etc. The MLIP package (short for Machine-Learning Interatomic Potentials) is available at https://mlip.skoltech.ru/download/.

引用

页数：19

共 83 条

[21] High-Dimensional Neural Network Potentials for Organic Reactions and an Improved Training Algorithm [J].

Gastegger, Michael ;

Marquetand, Philipp .

JOURNAL OF CHEMICAL THEORY AND COMPUTATION, 2015, 11 (05) :2187-2198

[22] USPEX - Evolutionary crystal structure prediction [J].

Glass, Colin W. ;

Oganov, Artem R. ;

Hansen, Nikolaus .

COMPUTER PHYSICS COMMUNICATIONS, 2006, 175 (11-12) :713-720

[23]

Goreinov S.A., 2010, Matrix Methods: Theory, Algorithms and Applications: Dedicated to the Memory of Gene Golub, P247, DOI 10.1142/97898128360210015

[24] Ab initio vibrational free energies including anharmonicity for multicomponent alloys [J].

Grabowski, Blazej ;

Ikeda, Yuji ;

Srinivasan, Prashanth ;

Kormann, Fritz ;

Freysoldt, Christoph ;

Duff, Andrew Ian ;

Shapeev, Alexander ;

Neugebauer, Joerg .

NPJ COMPUTATIONAL MATERIALS, 2019, 5 (1)

[25] Symmetry-Adapted Machine Learning for Tensorial Properties of Atomistic Systems [J].

Grisafi, Andrea ;

Wilkins, David M. ;

Csanyi, Gabor ;

Ceriotti, Michele .

PHYSICAL REVIEW LETTERS, 2018, 120 (03)

[26] Accelerating high-throughput searches for new alloys with active learning of interatomic potentials [J].

Gubaev, Konstantin ;

Podryabinkin, Evgeny, V ;

Hart, Gus L. W. ;

Shapeev, Alexander, V .

COMPUTATIONAL MATERIALS SCIENCE, 2019, 156 :148-156

[27] Machine learning of molecular properties: Locality and active learning [J].

Gubaev, Konstantin ;

Podryabinkin, Evgeny V. ;

Shapeev, Alexander V. .

JOURNAL OF CHEMICAL PHYSICS, 2018, 148 (24)

[28] Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space [J].

Hansen, Katja ;

Biegler, Franziska ;

Ramakrishnan, Raghunathan ;

Pronobis, Wiktor ;

von Lilienfeld, O. Anatole ;

Mueller, Klaus-Robert ;

Tkatchenko, Alexandre .

JOURNAL OF PHYSICAL CHEMISTRY LETTERS, 2015, 6 (12) :2326-2331

[29] Fast, accurate, and transferable many-body interatomic potentials by symbolic regression [J].

Hernandez, Alberto ;

Balasubramanian, Adarsh ;

Yuan, Fenglin ;

Mason, Simon A. M. ;

Mueller, Tim .

NPJ COMPUTATIONAL MATERIALS, 2019, 5 (1)

[30] Communication: Understanding molecular representations in machine learning: The role of uniqueness and target similarity [J].

Huang, Bing ;

von Lilienfeld, O. Anatole .

JOURNAL OF CHEMICAL PHYSICS, 2016, 145 (16)

← 1 2 3 4 5 6 7 8 9 →