ACEpotentials.jl: A Julia Implementation of the Atomic Cluster Expansion

TL;DR

ACEpotentials.jl is a Julia package that constructs accurate, efficient, and systematically improvable interatomic potentials using the Atomic Cluster Expansion, enabling rapid evaluation and active learning in atomistic simulations.

Contribution

It introduces ACEpotentials.jl, a Julia implementation of the Atomic Cluster Expansion for interatomic potentials, emphasizing simplicity, interpretability, and robustness.

Findings

Demonstrates high performance on atomistic modeling tasks

Shows systematic improvability and data efficiency

Facilitates active learning with Bayesian techniques

Abstract

We introduce ACEpotentials.jl, a Julia-language software package that constructs interatomic potentials from quantum mechanical reference data using the Atomic Cluster Expansion (Drautz, 2019). As the latter provides a complete description of atomic environments, including invariance to overall translation and rotation as well as permutation of like atoms, the resulting potentials are systematically improvable and data efficient. Furthermore, the descriptor's expressiveness enables use of a linear model, facilitating rapid evaluation and straightforward application of Bayesian techniques for active learning. We summarize the capabilities of ACEpotentials.jl and demonstrate its strengths (simplicity, interpretability, robustness, performance) on a selection of prototypical atomistic modelling workflows.