Conceptual and practical bases for the high accuracy of machine learning interatomic potentials: Application to elemental titanium

Akira Takahashi, Atsuto Seko, and Isao Tanaka

Phys. Rev. Materials 1, 063801 – Published 3 November 2017

Abstract

Machine learning interatomic potentials (MLIPs) based on a large data set obtained by density functional theory calculation have been developed recently. This study gives both conceptual and practical bases for the high accuracy of MLIPs, although MLIPs have been considered to be simply an accurate black-box description of atomic energy. We also construct the most accurate MLIP of elemental Ti ever reported using a linearized MLIP framework and many angular-dependent descriptors, which also corresponds to a generalization of the modified embedded atom method potential.

Received 12 August 2017

DOI:https://doi.org/10.1103/PhysRevMaterials.1.063801

Physics Subject Headings (PhySH)

Interatomic & molecular potentials Machine learning Potential energy surfaces

Quantum Information, Science & TechnologyAtomic, Molecular & OpticalGeneral PhysicsCondensed Matter, Materials & Applied Physics

Authors & Affiliations

Akira Takahashi^1,*, Atsuto Seko^1,2,3,4,†, and Isao Tanaka^1,2,3,5

¹Department of Materials Science and Engineering, Kyoto University, Kyoto 606-8501, Japan
²Center for Elements Strategy Initiative for Structure Materials (ESISM), Kyoto University, Kyoto 606-8501, Japan
³Center for Materials Research by Information Integration, National Institute for Materials Science, Tsukuba 305-0047, Japan
⁴JST, PRESTO, Kawaguchi 332-0012, Japan
⁵Nanostructures Research Laboratory, Japan Fine Ceramics Center, Nagoya 456-8587, Japan