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.
| Original language | English |
|---|---|
| Publisher | arXiv |
| DOIs | |
| Publication status | Published - 7 Sept 2023 |
Bibliographical note
Acknowledgments:GC acknowledges support from EPSRC grant EP/X035956/1. CO, AR and TJ were supported by NSERC Discovery Grant GR019381 and NFRF Exploration Grant GR022937. WB was supported by US AFRL grant FA8655-21-1-7010. C vd O and GC acknowledge ARCHER2 for which access was obtained via the UKCP consortium and funded by EPSRC grant EP/P022065/1. NB was supported by the U. S. Office of Naval Research through the U. S. Naval Research Laboratory’s fundamental research base program. EG acknowledges support from the EPSRC Centre for Doctoral Training in Automated Chemical Synthesis Enabled by Digital Molecular Technologies with grant reference EP/S024220/1. WCW was supported by the Schmidt Science Fellows in partnership with the Rhodes Trust, and additionally acknowledges support from EPSRC (Grant EP/V062654/1). JRK and CO acknowledge funding from the Leverhulme Trust under grant RPG-2017-191 and the EPSRC under grant EP/R043612/1. JRK, JPD and GC acknowledge support from the NOMAD Centre of Excellence funded by the European Commission under grant agreement 951786. JRK acknowledges support from the EPSRC under grants EP/P002188 and EP/R012474/1. This work was performed using resources provided by the Cambridge Service for Data Driven Discovery (CSD3) operated by the University of Cambridge Research Computing Service (www.csd3.cam.ac.uk), provided by Dell EMC and Intel using Tier-2 funding from the EPSRC (capital grant EP/T022159/1), and DiRAC funding from the STFC (www.dirac.ac.uk). Further computing facilities were provided by the Scientific Computing Research Technology Platform of the University of Warwick.