Automatic Selection of Atomic Fingerprints and Reference Configurations for Machine-Learning Potentials

TL;DR

This paper introduces an automatic method for selecting optimal atomic fingerprints and reference configurations to improve the accuracy and efficiency of machine-learning potentials in molecular modeling.

Contribution

It proposes a data-driven protocol for selecting descriptors from large pools, enhancing neural network potentials and Gaussian process regression applications.

Findings

Improved neural network potentials for water and Al-Mg-Si alloy.

Enhanced Gaussian process regression for organic molecule energies.

Potential to accelerate potential evaluation by orders of magnitude.

Abstract

Machine learning of atomic-scale properties is revolutionizing molecular modelling, making it possible to evaluate inter-atomic potentials with first-principles accuracy, at a fraction of the costs. The accuracy, speed and reliability of machine-learning potentials, however, depends strongly on the way atomic configurations are represented, i.e. the choice of descriptors used as input for the machine learning method. The raw Cartesian coordinates are typically transformed in "fingerprints", or "symmetry functions", that are designed to encode, in addition to the structure, important properties of the potential-energy surface like its invariances with respect to rotation, translation and permutation of like atoms. Here we discuss automatic protocols to select a number of fingerprints out of a large pool of candidates, based on the correlations that are intrinsic to the training data. This procedure can greatly simplify the construction of neural network potentials that strike the best balance between accuracy and computational efficiency, and has the potential to accelerate by orders of magnitude the evaluation of Gaussian Approximation Potentials based on the Smooth Overlap of Atomic Positions kernel. We present applications to the construction of neural network potentials for water and for an Al-Mg-Si alloy, and to the prediction of the formation energies of small organic molecules using Gaussian process regression.