Accelerating Atomistic Simulations with Piecewise Machine Learned Ab Initio Potentials at Classical Force Field-like Cost

Yaolong Zhang,Ce Hu,Bin Jiang

DOI: https://doi.org/10.1039/D0CP05089J

2020-11-17

Abstract:Machine learning methods have nowadays become easy-to-use tools for constructing high-dimensional interatomic potentials with ab initio accuracy. Although machine learned interatomic potentials are generally orders of magnitude faster than first-principles calculations, they remain much slower than classical force fields, at the price of using more complex structural descriptors. To bridge this efficiency gap, we propose an embedded atom neural network approach with simple piecewise switching function based descriptors, resulting in a favorable linear scaling with the number of neighbor atoms. Numerical examples validate that this piecewise machine learning model can be over an order of magnitude faster than various popular machine learned potentials with comparable accuracy for both metallic and covalent materials, approaching the speed of the fastest embedded atom method (i.e. several {\mu}s/atom per CPU core). The extreme efficiency of this approach promises its potential in first-principles atomistic simulations of very large systems and/or in long timescale.

Computational Physics,Chemical Physics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to improve the computational efficiency of machine - learning inter - atomic potential functions (MLIPs) while maintaining high accuracy, making it close to the speed of classical force fields (CFFs). Specifically, the authors propose an embedded - atom neural network (EANN) method based on piecewise - switching - function descriptors, called the piecewise - embedded - atom neural network (PEANN), to reduce the computational cost, especially when dealing with a large number of neighboring atoms. This method aims to narrow the efficiency gap between traditional MLIPs and CFFs by simplifying the calculation of structural descriptors, thereby achieving efficient first - principles atomic simulations for large - scale systems or long - time - scale simulations. The paper mentions that although existing machine - learning inter - atomic potential functions are superior to classical force fields in terms of accuracy, their computational cost is still high, usually at least two orders of magnitude slower than the fastest embedded - atom method (EAM) force field. Therefore, the goal of the paper is to develop a method that can maintain high accuracy and significantly improve computational efficiency, so that it can be applied to larger - scale systems and longer - time - scale simulations.

Accelerating Atomistic Simulations with Piecewise Machine Learned Ab Initio Potentials at Classical Force Field-like Cost

Accelerating atomistic simulations with piecewise machine-learned ab Initio potentials at a classical force field-like cost

Accurate interatomic force field for molecular dynamics simulation by hybridizing classical and machine learning potentials

Efficient Machine Learning Force Field for Large-Scale Molecular Simulations of Organic Systems

Indirect Learning of Interatomic Potentials for Accelerated Materials Simulations

Ab-Initio Molecular Dynamics Acceleration Scheme with an Adaptive Machine Learning Framework

Learning Interatomic Potentials at Multiple Scales

In operando active learning of interatomic interaction during large-scale simulations

Extending the atomic decomposition and many-body representation, a chemistry-motivated monomer-centered approach for machine learning potentials

Machine‐learning‐based Interatomic Potentials for Advanced Manufacturing

Molecular Dynamics with On-the-Fly Machine Learning of Quantum-Mechanical Forces

Towards Exact Molecular Dynamics Simulations with Machine-Learned Force Fields

Adaptive-precision potentials for large-scale atomistic simulations

Strategies for the Construction of Machine-Learning Potentials for Accurate and Efficient Atomic-Scale Simulations

Embedded Atom Neural Network Potentials: Efficient and Accurate Machine Learning with a Physically Inspired Representation

Fast and accurate machine-learned interatomic potentials for large-scale simulations of Cu, Al and Ni

Accuracy, Performance, and Transferability of Interparticle Potentials for Al–Cu Alloys: Comparison of Embedded Atom and Deep Machine Learning Models

Introduction to machine learning potentials for atomistic simulations

NNP/MM: Accelerating Molecular Dynamics Simulations with Machine Learning Potentials and Molecular Mechanics

Machine Learning Interatomic Potentials and Long-Range Physics

Accelerating high-throughput searches for new alloys with active learning of interatomic potentials