Repository logo
 

Hyperactive learning for data-driven interatomic potentials

Accepted version
Peer-reviewed

Type

Article

Change log

Authors

van der Oord, Cas 
Csanyi, Gabor 
Sachs, Matthias 
Kovacs, David Peter 
Ortner, Christoph 

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) in order to drive the system towards unseen 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.

Description

Keywords

40 Engineering, 34 Chemical Sciences, 3406 Physical Chemistry

Journal Title

npj Computational Materials

Conference Name

Journal ISSN

2057-3960

Volume Title

Publisher

Nature Portfolio
Sponsorship
Engineering and Physical Sciences Research Council (1971218)
Engineering and Physical Sciences Research Council (EP/K014560/1)
Relationships
Is previous version of: