Simple machine-learned interatomic potentials for complex alloys

TL;DR

This paper demonstrates that simple low-dimensional descriptors can efficiently develop accurate machine-learning interatomic potentials for complex multi-element alloys, outperforming more complex descriptors in data efficiency and speed.

Contribution

It introduces a computationally fast tabGAP potential using simple descriptors for multi-element alloys, improving accuracy and efficiency over complex descriptors.

Findings

Low-dimensional descriptors achieve meV/atom accuracy.

Simple descriptors outperform SOAP in data efficiency.

Developed a fast tabGAP potential for Mo-Nb-Ta-V-W alloys.

Abstract

Developing data-driven machine-learning interatomic potentials for materials containing many elements becomes increasingly challenging due to the vast configuration space that must be sampled by the training data. We study the learning rates and achievable accuracy of machine-learning interatomic potentials for many-element alloys with different combinations of descriptors for the local atomic environments. We show that for a five-element alloy system, potentials using simple low-dimensional descriptors can reach meV/atom-accuracy with modestly sized training datasets, significantly outperforming the high-dimensional SOAP descriptor in data efficiency, accuracy, and speed. In particular, we develop a computationally fast machine-learned and tabulated Gaussian approximation potential (tabGAP) for Mo-Nb-Ta-V-W alloys with a combination of two-body, three-body, and a new simple scalar many-body density descriptor based on the embedded atom method.