Strategies for the Construction of Machine-Learning Potentials for Accurate and Efficient Atomic-Scale Simulations

TL;DR

This paper provides a comprehensive tutorial on constructing machine-learning neural network potentials for atomistic simulations, covering data collection, model training, validation, and refinement, to help researchers adopt these methods more easily.

Contribution

It offers practical strategies and best practices for building neural network potentials, including discussions on active learning and delta learning, tailored for chemists and materials scientists.

Findings

Guidelines for data collection and model selection.

Examples demonstrating training and validation processes.

Discussion on active learning and delta learning techniques.

Abstract

Recent advances in machine-learning interatomic potentials have enabled the efficient modeling of complex atomistic systems with an accuracy that is comparable to that of conventional quantum mechanics based methods. At the same time, the construction of new machine-learning potentials can seem a daunting task, as it involves data-science techniques that are not yet common in chemistry and materials science. Here, we provide a tutorial-style overview of strategies and best practices for the construction of artificial neural network (ANN) potentials. We illustrate the most important aspects of (i) data collection, (ii) model selection, (iii) training and validation, and (iv) testing and refinement of ANN potentials on the basis of practical examples. Current research in the areas of active learning and delta learning are also discussed in the context of ANN potentials. This tutorial review aims at equipping computational chemists and materials scientists with the required background knowledge for ANN potential construction and application, with the intention to accelerate the adoption of the method, so that it can facilitate exciting research that would otherwise be challenging with conventional strategies.