Active learning strategies for atomic cluster expansion models

TL;DR

This paper compares two uncertainty estimation methods for atomic cluster expansion models, highlighting the efficiency of the D-optimality approach and its potential for automating data selection and discovering rare atomic configurations.

Contribution

It introduces and evaluates active learning strategies for ACE models, emphasizing the efficiency of the D-optimality criterion in uncertainty estimation and data exploration.

Findings

D-optimality based extrapolation grade is more computationally efficient.

Both methods show comparable prediction accuracy.

Active learning enables exploration of new atomic structures.

Abstract

The atomic cluster expansion (ACE) was proposed recently as a new class of data-driven interatomic potentials with a formally complete basis set. Since the development of any interatomic potential requires a careful selection of training data and thorough validation, an automation of the construction of the training dataset as well as an indication of a model's uncertainty are highly desirable. In this work, we compare the performance of two approaches for uncertainty indication of ACE models based on the D-optimality criterion and ensemble learning. While both approaches show comparable predictions, the extrapolation grade based on the D-optimality (MaxVol algorithm) is more computationally efficient. In addition, the extrapolation grade indicator enables an active exploration of new structures, opening the way to the automated discovery of rare-event configurations. We demonstrate that active learning is also applicable to explore local atomic environments from large-scale MD simulations.