Interatomic potential construction with self-learning and adaptive database

doi:10.1103/PhysRevMaterials.1.053801

Interatomic potential construction with self-learning and adaptive database

2017-10-02

38

PRE

AI

DOI：10.1103/PhysRevMaterials.1.053801

Original

Original request for help

Save

Abstract

En 中文

We have developed an automatic machine learning potential (MLP) construction scheme, the self-learning and adaptive database (SLAD). The sample structures for training are collected by the molecular dynamics simulations using the MLP itself with the aid of the spilling factor for simultaneous error estimation. The utility of the SLAD is demonstrated by applying it for the solid-state ionic conductor, Li2B12H12. Starting from the low-temperature alpha-phase structure, the MLP is successfully developed, for which the total number of density functional calculations required is only 84. The constructed MLP reproduces well the phase transition to the high-temperature beta phase accompanied by an abrupt lattice expansion, promotion of Li diffusion, and orientational disordering of B12H12 complexes. The SLAD allows us the efficient construction of MLPs for complicated systems.

Keywords:

GENERALIZED GRADIENT APPROXIMATION

SOLID ELECTROLYTES

DYNAMICS

LI2B12H12

Interatomic potential construction with self-learning and adaptive database

Abstract

Journal

Organization