An automated approach for developing neural network interatomic potentials with FLAME

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.

[1]  K-R Müller,et al.  SchNetPack: A Deep Learning Toolbox For Atomistic Systems. , 2018, Journal of chemical theory and computation.

[2]  David W Toth,et al.  The TensorMol-0.1 model chemistry: a neural network augmented with long-range physics , 2017, Chemical science.

[3]  K. Sumathy,et al.  A review and recent developments in photocatalytic water-splitting using TiO2 for hydrogen production , 2007 .

[4]  S. Goedecker.  Minima hopping: an efficient search method for the global minimum of the potential energy surface of complex molecular systems. , 2004, The Journal of chemical physics.

[5]  Satoshi Watanabe,et al.  Study of Li atom diffusion in amorphous Li3PO4 with neural network potential. , 2017, The Journal of chemical physics.

[6]  Christian Trott,et al.  Spectral neighbor analysis method for automated generation of quantum-accurate interatomic potentials , 2014, J. Comput. Phys..

[7]  J S Smith,et al.  ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost , 2016, Chemical science.

[8]  G. Henkelman,et al.  Alkali Atoms Diffusion Mechanism in CuInSe2 Explained by Kinetic Monte Carlo Simulations , 2019, Advanced Theory and Simulations.

[9]  Mario Valle,et al.  How to quantify energy landscapes of solids. , 2009, The Journal of chemical physics.

[10]  Burke,et al.  Generalized Gradient Approximation Made Simple. , 1996, Physical review letters.

[11]  E Weinan,et al.  Active Learning of Uniformly Accurate Inter-atomic Potentials for Materials Simulation , 2018, Physical Review Materials.

[12]  I. Tanaka,et al.  First principles phonon calculations in materials science , 2015, 1506.08498.

[13]  Alexander Kaiser,et al.  A neural network interface for DL_POLY and its application to liquid water , 2018, Molecular Simulation.

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

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

[16]  S. Goedecker,et al.  High accuracy and transferability of a neural network potential through charge equilibration for calcium fluoride , 2017 .

[17]  E Weinan,et al.  Deep Potential Molecular Dynamics: a scalable model with the accuracy of quantum mechanics , 2017, Physical review letters.

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

[19]  Nongnuch Artrith,et al.  High-dimensional neural-network potentials for multicomponent systems: Applications to zinc oxide , 2011 .

[20]  T. Morawietz,et al.  How van der Waals interactions determine the unique properties of water , 2016, Proceedings of the National Academy of Sciences.

[21]  G. A. Slack,et al.  Nonmetallic crystals with high thermal conductivity , 1973 .

[22]  D. Donadio,et al.  Transferability of neural network potentials for varying stoichiometry: Phonons and thermal conductivity of MnxGey compounds , 2020, Journal of Applied Physics.

[23]  M. Mohammadizadeh,et al.  First-principles elastic and thermal properties of TiO2: a phonon approach , 2010, Journal of physics. Condensed matter : an Institute of Physics journal.

[24]  Maximilian Amsler,et al.  FLAME: A library of atomistic modeling environments , 2019, Comput. Phys. Commun..

[25]  Wei Chen,et al.  FireWorks: a dynamic workflow system designed for high‐throughput applications , 2015, Concurr. Comput. Pract. Exp..

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

[27]  G. Madsen,et al.  Comparing the performance of LDA and GGA functionals in predicting the lattice thermal conductivity of III-V semiconductor materials in the zincblende structure: The cases of AlAs and BAs , 2018, Computational Materials Science.

[28]  Jörg Behler,et al.  Structure and Dynamics of the Liquid–Water/Zinc-Oxide Interface from Machine Learning Potential Simulations , 2019, The Journal of Physical Chemistry C.

[29]  Rustam Z. Khaliullin,et al.  Microscopic origins of the anomalous melting behavior of sodium under high pressure. , 2011, Physical review letters.

[30]  Stefan Goedecker,et al.  Crystal structure prediction using the minima hopping method. , 2010, The Journal of chemical physics.

[31]  E Weinan,et al.  DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics , 2017, Comput. Phys. Commun..

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

[33]  S. Sharma,et al.  Perspectives of chalcopyrite-based CIGSe thin-film solar cell: a review , 2020, Journal of Materials Science: Materials in Electronics.

[34]  S. Ghasemi,et al.  Optimized symmetry functions for machine-learning interatomic potentials of multicomponent systems. , 2018, The Journal of chemical physics.

[35]  G. Henkelman,et al.  A climbing image nudged elastic band method for finding saddle points and minimum energy paths , 2000 .

[36]  Wu Li,et al.  ShengBTE: A solver of the Boltzmann transport equation for phonons , 2014, Comput. Phys. Commun..

[37]  Blöchl,et al.  Projector augmented-wave method. , 1994, Physical review. B, Condensed matter.

[38]  Jörg Behler,et al.  Proton-Transfer Mechanisms at the Water-ZnO Interface: The Role of Presolvation. , 2017, The journal of physical chemistry letters.

[39]  Woong Kim,et al.  Cu(In,Ga)(S,Se)2 Photocathodes with a Grown-In CuxS Catalyst for Solar Water Splitting , 2019, ACS Energy Letters.

[40]  Nicholas Lubbers,et al.  Automated discovery of a robust interatomic potential for aluminum , 2020, Nature Communications.

[41]  Yi He,et al.  A Deep-Learning Potential for Crystalline and Amorphous Li–Si Alloys , 2020 .

[42]  V. Korolev,et al.  Neural Network Based Modeling of Grain Boundary Complexions Localized in Simple Symmetric Tilt Boundaries Σ3 (111) and Σ5 (210) , 2020, Colloid Journal.

[43]  Kresse,et al.  Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. , 1996, Physical review. B, Condensed matter.

[44]  Rustam Z. Khaliullin,et al.  Graphite-diamond phase coexistence study employing a neural-network mapping of the ab initio potential energy surface , 2010 .

[45]  S. Goedecker,et al.  Two-Dimensional Hexagonal Sheet of TiO2 , 2017 .

[46]  A. Fujishima,et al.  Electrochemical Photolysis of Water at a Semiconductor Electrode , 1972, Nature.

[47]  Anubhav Jain,et al.  Data mined ionic substitutions for the discovery of new compounds. , 2011, Inorganic chemistry.

[48]  Thermal Conductivity of Rutile and Anatase TiO2 from First-Principles , 2019 .

[49]  G. Kresse,et al.  Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set , 1996 .

[50]  Stefan Goedecker,et al.  Interatomic potentials for ionic systems with density functional accuracy based on charge densities obtained by a neural network , 2015, 1501.07344.

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

[52]  Ibrahim Khaled,et al.  Measuring transferability issues in machine-learning force fields: the example of gold–iron interactions with linearized potentials , 2019, Mach. Learn. Sci. Technol..

[53]  Alexandre Tkatchenko,et al.  Non-covalent interactions across organic and biological subsets of chemical space: Physics-based potentials parametrized from machine learning. , 2017, The Journal of chemical physics.

[54]  Anubhav Jain,et al.  The Materials Application Programming Interface (API): A simple, flexible and efficient API for materials data based on REpresentational State Transfer (REST) principles , 2015 .

[55]  Jimmy C. Yu,et al.  Cu(In,Ga)Se2 for selective and efficient photoelectrochemical conversion of CO2 into CO , 2020 .

[56]  Alexander V. Shapeev,et al.  Active learning of linearly parametrized interatomic potentials , 2016, 1611.09346.

[57]  Jörg Behler,et al.  From Molecular Fragments to the Bulk: Development of a Neural Network Potential for MOF-5. , 2019, Journal of chemical theory and computation.

[58]  J. Nørskov,et al.  Solvent–Adsorbate Interactions and Adsorbate-Specific Solvent Structure in Carbon Dioxide Reduction on a Stepped Cu Surface , 2019, The Journal of Physical Chemistry C.

[59]  S. Sahoo,et al.  In silico investigation of Cu(In,Ga)Se2-based solar cells. , 2020, Physical chemistry chemical physics : PCCP.

[60]  Matthew Horton,et al.  Atomate: A high-level interface to generate, execute, and analyze computational materials science workflows , 2017 .

[61]  Li Li,et al.  Bypassing the Kohn-Sham equations with machine learning , 2016, Nature Communications.

[62]  Jörg Behler,et al.  Neural network molecular dynamics simulations of solid-liquid interfaces: water at low-index copper surfaces. , 2016, Physical chemistry chemical physics : PCCP.

[63]  Noam Bernstein,et al.  Machine Learning a General-Purpose Interatomic Potential for Silicon , 2018, Physical Review X.

[64]  Thomas D. Kuhne,et al.  Ab initio quality neural-network potential for sodium , 2010, 1002.2879.

[65]  Anubhav Jain,et al.  Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysis , 2012 .

[66]  S. Ghasemi,et al.  Energy landscape of ZnO clusters and low-density polymorphs , 2017 .

[67]  K. Müller,et al.  Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space , 2015, The journal of physical chemistry letters.

[68]  Kristin A. Persson,et al.  Commentary: The Materials Project: A materials genome approach to accelerating materials innovation , 2013 .

[69]  Xiaobo Chen,et al.  Titanium dioxide-based nanomaterials for photocatalytic fuel generations. , 2014, Chemical reviews.

[70]  Hafner,et al.  Ab initio molecular dynamics for liquid metals. , 1995, Physical review. B, Condensed matter.

[71]  Michele Parrinello,et al.  Nucleation mechanism for the direct graphite-to-diamond phase transition. , 2011, Nature materials.

[72]  S. Goedecker,et al.  Surface reconstructions and premelting of the (100) CaF2 surface. , 2019, Physical chemistry chemical physics : PCCP.

[73]  Yuval Elbaz,et al.  Modeling Diffusion in Functional Materials: From Density Functional Theory to Artificial Intelligence , 2019, Advanced Functional Materials.