Learning Smooth and Expressive Interatomic Potentials for Physical Property Prediction

TL;DR

This paper introduces a new machine learning interatomic potential model, eSEN, that emphasizes energy conservation during simulations, leading to improved physical property predictions across various tasks.

Contribution

The paper proposes a practical evaluation method for MLIPs based on energy conservation in simulations and develops the eSEN model that achieves state-of-the-art results.

Findings

eSEN outperforms previous models on physical property prediction tasks.

Energy conservation during simulations correlates with better property prediction.

Identified model choices that affect energy conservation and prediction accuracy.

Abstract

Machine learning interatomic potentials (MLIPs) have become increasingly effective at approximating quantum mechanical calculations at a fraction of the computational cost. However, lower errors on held out test sets do not always translate to improved results on downstream physical property prediction tasks. In this paper, we propose testing MLIPs on their practical ability to conserve energy during molecular dynamic simulations. If passed, improved correlations are found between test errors and their performance on physical property prediction tasks. We identify choices which may lead to models failing this test, and use these observations to improve upon highly-expressive models. The resulting model, eSEN, provides state-of-the-art results on a range of physical property prediction tasks, including materials stability prediction, thermal conductivity prediction, and phonon calculations.