A Novel Approach to Describe Chemical Environments in High Dimensional Neural Network Potentials

Abstract

A central concern of molecular dynamics simulations are 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 SOAP descriptors currently in the literature.

arXiv comment23 Pages, 5 figures, 2 tables, journal article

References in corpus (12)

• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• On representing chemical environments
• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Perspective on density functional theory
• Comparing molecules and solids across structural and alchemical space
• Machine-learning based interatomic potential for amorphous carbon
• Finding Density Functionals with Machine Learning
• How van der Waals interactions determine the unique properties of water
• Achieving DFT accuracy with a machine-learning interatomic potential: thermomechanics and defects in bcc ferromagnetic iron
• Discovering the Building Blocks of Atomic Systems using Machine Learning
• Effects of Cutoff Functions of Tersoff Potentials on Molecular Dynamics Simulations of Thermal Transport
• The classical molecular dynamics simulation of graphene on Ru(0001) using a fitted Tersoff interface potential

Cited by in corpus (17)

• A Universal Graph Deep Learning Interatomic Potential for the Periodic Table
• Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
• Representations of molecules and materials for interpolation of quantum-mechanical simulations via machine learning
• Gaussian Moments as Physically Inspired Molecular Descriptors for Accurate and Scalable Machine Learning Potentials
• Strategies for the Construction of Machine-Learning Potentials for Accurate and Efficient Atomic-Scale Simulations
• An assessment of the structural resolution of various fingerprints commonly used in machine learning
• Predicting Properties of Periodic Systems from Cluster Data: A Case Study of Liquid Water
• Fast and Sample-Efficient Interatomic Neural Network Potentials for Molecules and Materials Based on Gaussian Moments
• Manifolds of quasi-constant SOAP and ACSF fingerprints and the resulting failure to machine learn four-body interactions
• Lightweight and Effective Tensor Sensitivity for Atomistic Neural Networks
• Compact atomic descriptors enable accurate predictions via linear models
• Hydrogen-induced degradation dynamics in silicon heterojunction solar cells via machine learning
• Transferable empirical pseudopotenials from machine learning
• Apax: A Flexible and Performant Framework For The Development of Machine-Learned Interatomic Potentials
• Benchmarking Structural Evolution Methods for Training of Machine Learned Interatomic Potentials
• Phase transitions of correlated systems from graph neural networks with quantum embedding techniques
• REANN: A PyTorch-based End-to-End Multi-functional Deep Neural Network Package for Molecular, Reactive and Periodic Systems