Machine Learning Interatomic Potentials for Million-Atom Simulations of Multicomponent Alloys

TL;DR

This paper compares two advanced machine learning interatomic potential frameworks, NEP and GRACE, for simulating complex alloys, highlighting NEP's superior speed and ensemble-based uncertainty quantification for large-scale dynamic simulations.

Contribution

The study evaluates and contrasts NEP and GRACE MLIP frameworks for multicomponent alloys, demonstrating NEP's high inference speed and effective uncertainty quantification in large-scale simulations.

Findings

GRACE shows higher training efficiency and slightly better accuracy.

NEP achieves 60-fold higher inference speed, suitable for million-atom simulations.

Ensemble-based uncertainty correlates well with model error, aiding reliable simulations.

Abstract

Machine learning interatomic potentials (MLIPs) with broad chemical flexibility are important for atomistic simulations of compositionally complex materials such as high-entropy alloys. Here, we study two state-of-the-art MLIP frameworks, the neuroevolution potential (NEP) and the graph atomic cluster expansion (GRACE), for 16 elemental metals and multicomponent alloys. GRACE potential with Finnis-Sinclair type shows substantially higher training efficiency and consistently, though only slightly, better accuracy for mechanical properties, thermal stability, and chemical extrapolation. In contrast, NEP achieves an approximately 60-fold higher inference speed, making it attractive for million-atom molecular dynamics simulations. We further examine uncertainty quantification strategies and find that ensemble-based uncertainty correlates robustly with model error, whereas D-optimality is less reliable for the systems considered here. Large-scale nonequilibrium molecular dynamics simulations of shock propagation further show that NEP, combined with ensemble-based uncertainty quantification, enables efficient and reliable simulations under extreme dynamic conditions.