The Importance of Learning without Constraints: Reevaluating Benchmarks for Invariant and Equivariant Features of Machine Learning Potentials in Generating Free Energy Landscapes

TL;DR

This paper reevaluates benchmarks for machine-learned interatomic potentials, emphasizing the importance of unconstrained, unbiased datasets and enhanced sampling techniques for stable, accurate, and extrapolatable molecular simulations.

Contribution

It highlights the impact of training dataset characteristics on MILP stability and accuracy, proposing methods to improve extrapolation and simulation reliability.

Findings

Unconstrained datasets improve MILP stability.

Enhanced sampling enhances energy prediction accuracy.

Proper symmetry considerations aid safe extrapolation.

Abstract

Machine-learned interatomic potentials (MILPs) are rapidly gaining interest for molecular modeling, as they provide a balance between quantum-mechanical level descriptions of atomic interactions and reasonable computational efficiency. However, questions remain regarding the stability of simulations using these potentials, as well as the extent to which the learned potential energy function can be extrapolated safely. Past studies have reported challenges encountered when MILPs are applied to classical benchmark systems. In this work, we show that some of these challenges are related to the characteristics of the training datasets, particularly the inclusion of rigid constraints. We demonstrate that long stability in simulations with MILPs can be achieved by generating unconstrained datasets using unbiased classical simulations if the fast modes are correctly sampled. Additionally, we emphasize that in order to achieve precise energy predictions, it is important to resort to enhanced sampling techniques for dataset generation, and we demonstrate that safe extrapolation of MILPs depends on judicious choices related to the system's underlying free energy landscape and the symmetry features embedded within the machine learning models.