Environment-adaptive machine learning potentials
Abstract
The development of interatomic potentials that can accurately capture a wide range of physical phenomena and diverse environments is of significant interest, but it presents a formidable challenge. This challenge arises from the numerous structural forms, multiple phases, complex intramolecular and intermolecular interactions, and varying external conditions. In this paper, we present a method to construct environment-adaptive interatomic potentials by adapting to the local atomic environment of each atom within a system. The collection of atomic environments of interest is partitioned into several clusters of atomic environments. Each cluster represents a distinctive local environment and is used to define a corresponding local potential. We introduce a many-body many-potential expansion to smoothly blend these local potentials to ensure global continuity of the potential energy surface. This is achieved by computing the probability functions that determine the likelihood of an atom belonging to each cluster. We apply the environment-adaptive machine learning potentials to predict observable properties for Ta element and InP compound, and compare them with density functional theory calculations.
AI-Generated Overview
-
Research Focus: The study focuses on developing environment-adaptive machine learning potentials (EAML) for accurately capturing the potential energy surface of atomic systems across diverse physical phenomena and environments.
-
Methodology: The authors propose a systematic approach where local atomic environments are categorized into clusters, each associated with distinct local potentials. A many-body many-potential expansion is introduced to blend these local potentials, ensuring global continuity of the energy surface through probability functions.
-
Results: The application of EAML potentials to predict observable properties of Tantalum (Ta) and Indium Phosphide (InP) shows significant accuracy improvements over existing methods. The results demonstrate better performance against density functional theory calculations for both elements.
-
Key Contribution(s): The introduction of proper orthogonal descriptors tailored to multi-element systems and the innovative blending of local potentials through a many-body framework are central contributions, enhancing the flexibility and accuracy of interatomic potentials.
-
Significance: EAML potentials can effectively adapt to varying atomic environments, providing reliable predictions for a wide range of structural forms and chemical states, which addresses critical limitations in traditional empirical potentials.
-
Broader Applications: The methodology has potential applications in various fields, including materials science and computational chemistry, where understanding interatomic forces in complex systems under different conditions is essential—facilitating accurate simulations in molecular dynamics and other computational techniques.