A novel approach to describe chemical environments in high-dimensional neural network potentials
Open Access
- 15 April 2019
- journal article
- conference paper
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 150 (15) , 154102
Abstract
A central concern of molecular dynamics simulations is the potential energy surfaces that govern atomic interactions. These hypersurfaces define the potential energy of the system and have generally been calculated using either predefined analytical formulas (classical) or quantum mechanical simulations (ab initio). The former can accurately reproduce only a selection of material properties, whereas the latter is restricted to short simulation times and small systems. Machine learning potentials have recently emerged as a third approach to model atomic interactions, and are purported to offer the accuracy of ab initio simulations with the speed of classical potentials. However, the performance of machine learning potentials depends crucially on the description of a local atomic environment. A set of invariant, orthogonal, and differentiable descriptors for an atomic environment is proposed, implemented in a neural network potential for solid-state silicon, and tested in molecular dynamics simulations. Neural networks using the proposed descriptors are found to outperform ones using the Behler–Parinello and smooth overlap of atomic position descriptors in the literature.Keywords
All Related Versions
This publication has 40 references indexed in Scilit:
- Effects of cutoff functions of Tersoff potentials on molecular dynamics simulations of thermal transportModelling and Simulation in Materials Science and Engineering, 2011
- Pressure‐induced phase transitions in silicon studied by neural network‐based metadynamics simulationsPhysica Status Solidi (b), 2008
- Generalized Neural-Network Representation of High-Dimensional Potential-Energy SurfacesPhysical Review Letters, 2007
- Representing high-dimensional potential-energy surfaces for reactions at surfaces by neural networksChemical Physics Letters, 2004
- The fitting of potential energy and transition moment functions using neural networks: transition probabilities in OH (A2Σ+→X2Π)Chemical Physics, 2004
- The fitting of potential energy surfaces using neural networks. Application to the study of the photodissociation processesChemical Physics Letters, 1998
- Fast Parallel Algorithms for Short-Range Molecular DynamicsJournal of Computational Physics, 1995
- Incorporation of Monovalent Ions in ZnO and Their Influence on Varistor DegradationJournal of the American Ceramic Society, 1993
- Learning representations by back-propagating errorsNature, 1986
- Computer simulation of local order in condensed phases of siliconPhysical Review B, 1985