Abstract
Data-driven interatomic potentials have emerged as a powerful tool for approximating ab initio potential energy surfaces. The most time-consuming step in creating these interatomic potentials is typically the generation of a suitable training database. To aid this process hyperactive learning (HAL), an accelerated active learning scheme, is presented as a method for rapid automated training database assembly. HAL adds a biasing term to a physically motivated sampler (e.g. molecular dynamics) driving atomic structures towards uncertainty in turn generating unseen or valuable training configurations. The proposed HAL framework is used to develop atomic cluster expansion (ACE) interatomic potentials for the AlSi10 alloy and polyethylene glycol (PEG) polymer starting from roughly a dozen initial configurations. The HAL generated ACE potentials are shown to be able to determine macroscopic properties, such as melting temperature and density, with close to experimental accuracy.
Original language | English |
---|---|
Article number | 168 |
Number of pages | 14 |
Journal | npj Computational Materials |
Volume | 9 |
Issue number | 1 |
DOIs | |
Publication status | Published - 13 Sept 2023 |
Bibliographical note
Funding Information:G.C. and C.v.d.O. acknowledge the support of UKCP grant number EP/K014560/1. C.v.d.O. would like to acknowledge the support of EPSRC (Project Reference: 1971218) and Dassault Systèmes UK. C.O. acknowledges support of the NSERC Discovery Grant (IDGR019381) and the NFRF Exploration Grant GR022937. The authors would also like to thank Ioan-Bogdan Magdău for discussions on modelling condensed phase polymers.
Publisher Copyright:
© 2023, Springer Nature Limited.
ASJC Scopus subject areas
- Modelling and Simulation
- General Materials Science
- Mechanics of Materials
- Computer Science Applications