Machine learning for atomic forces in a crystalline solid: Transferability to various temperatures

Recently, machine learning has emerged as an alternative, powerful approach for predicting quantum-mechanical properties of molecules and solids. Here, using kernel ridge regression and atomic fingerprints representing local environments of atoms, we trained a machine-learning model on a crystalline silicon system to directly predict the atomic forces at a wide range of temperatures. Our idea is to construct a machine-learning model using a quantum-mechanical dataset taken from canonical-ensemble simulations at a higher temperature, or an upper bound of the temperature range. With our model, the force prediction errors were about 2% or smaller with respect to the corresponding force ranges, in the temperature region between 300 K and 1650 K. We also verified the applicability to a larger system, ensuring the transferability with respect to system size.

[1]  R. Kondor,et al.  Gaussian approximation potentials: the accuracy of quantum mechanics, without the electrons. , 2009, Physical review letters.

[2]  G. G. Stokes.  "J." , 1890, The New Yale Book of Quotations.

[3]  V. Ozoliņš,et al.  Lattice anharmonicity and thermal conductivity from compressive sensing of first-principles calculations. , 2014, Physical review letters.

[4]  David R. Bowler,et al.  Recent progress in linear scaling ab initio electronic structure techniques , 2002 .

[5]  Andrew G. Glen,et al.  APPL , 2001 .

[6]  David R. Bowler,et al.  Recent progress with large‐scale ab initio calculations: the CONQUEST code , 2006 .

[7]  Richard J. Needs,et al.  Tests of the Harris energy functional , 1989 .

[8]  W. Kohn,et al.  Self-Consistent Equations Including Exchange and Correlation Effects , 1965 .

[9]  D R Bowler,et al.  Atomic force algorithms in density functional theory electronic-structure techniques based on local orbitals. , 2004, The Journal of chemical physics.

[10]  Y. Gohda,et al.  Anharmonic force constants extracted from first-principles molecular dynamics: applications to heat transfer simulations , 2014, Journal of physics. Condensed matter : an Institute of Physics journal.

[11]  Rampi Ramprasad,et al.  Learning scheme to predict atomic forces and accelerate materials simulations , 2015, 1505.02701.

[12]  J. Behler.  Neural network potential-energy surfaces in chemistry: a tool for large-scale simulations. , 2011, Physical chemistry chemical physics : PCCP.

[13]  Nongnuch Artrith,et al.  An implementation of artificial neural-network potentials for atomistic materials simulations: Performance for TiO2 , 2016 .

[14]  Alessandro De Vita,et al.  A framework for machine‐learning‐augmented multiscale atomistic simulations on parallel supercomputers , 2015 .

[15]  Taku Ohara,et al.  Thermal conductivity of silicon nanowire by nonequilibrium molecular dynamics simulations , 2009 .

[16]  Harris.  Simplified method for calculating the energy of weakly interacting fragments. , 1985, Physical review. B, Condensed matter.

[17]  Sebastian Volz,et al.  Molecular dynamics simulation of thermal conductivity of silicon nanowires , 1999 .

[18]  Michele Parrinello,et al.  Generalized neural-network representation of high-dimensional potential-energy surfaces. , 2007, Physical review letters.

[19]  S. Ju,et al.  Thermal conductivity of nanocrystalline silicon by direct molecular dynamics simulation , 2012 .

[20]  J. Behler.  Atom-centered symmetry functions for constructing high-dimensional neural network potentials. , 2011, The Journal of chemical physics.

[21]  F. Müller-Plathe.  A simple nonequilibrium molecular dynamics method for calculating the thermal conductivity , 1997 .

[22]  Matthias Rupp,et al.  Machine learning for quantum mechanics in a nutshell , 2015 .

[23]  Rampi Ramprasad,et al.  Adaptive machine learning framework to accelerate ab initio molecular dynamics , 2015 .

[24]  R. Kondor,et al.  On representing chemical environments , 2012, 1209.3140.

[25]  S. Tsuneyuki,et al.  Self-consistent phonon calculations of lattice dynamical properties in cubic SrTiO 3 with first-principles anharmonic force constants , 2015, 1506.01781.

[26]  Klaus-Robert Müller,et al.  Assessment and Validation of Machine Learning Methods for Predicting Molecular Atomization Energies. , 2013, Journal of chemical theory and computation.

[27]  M C Shaughnessy,et al.  Efficient Use of an Adapting Database of Ab Initio Calculations To Generate Accurate Newtonian Dynamics. , 2016, Journal of chemical theory and computation.

[28]  Sanguthevar Rajasekaran,et al.  Accelerating materials property predictions using machine learning , 2013, Scientific Reports.

[29]  Mark E. Tuckerman,et al.  Explicit reversible integrators for extended systems dynamics , 1996 .

[30]  M. Klein,et al.  Nosé-Hoover chains : the canonical ensemble via continuous dynamics , 1992 .

[31]  Martins,et al.  Efficient pseudopotentials for plane-wave calculations. , 1991, Physical review. B, Condensed matter.

[32]  Atsuto Seko,et al.  First-principles interatomic potentials for ten elemental metals via compressed sensing , 2015, 1505.03994.

[33]  K. Müller,et al.  Fast and accurate modeling of molecular atomization energies with machine learning. , 2011, Physical review letters.

[34]  Foulkes,et al.  Tight-binding models and density-functional theory. , 1989, Physical review. B, Condensed matter.

[35]  Lewis,et al.  Self-diffusion on low-index metallic surfaces: Ag and Au (100) and (111). , 1996, Physical review. B, Condensed matter.

[36]  Zhenwei Li,et al.  Molecular dynamics with on-the-fly machine learning of quantum-mechanical forces. , 2015, Physical review letters.

[37]  M. Rupp,et al.  Machine Learning for Quantum Mechanical Properties of Atoms in Molecules , 2015, 1505.00350.