Linear Jacobi-Legendre expansion of the charge density for machine learning-accelerated electronic structure calculations

Autor(es)

Focassio, Bruno ; Domina, Michelangelo ; Patil, Urvesh ; Fazzio, Adalberto ; Sanvito, Stefano

• Tipo de documento: Artigo Científico
• Fonte: NPJ COMPUTATIONAL MATERIALS; v. 9, n. 1, p. 10-pg., 2023-05-29.

Resumo

Kohn-Sham density functional theory (KS-DFT) is a powerful method to obtain key materials' properties, but the iterative solution of the KS equations is a numerically intensive task, which limits its application to complex systems. To address this issue, machine learning (ML) models can be used as surrogates to find the ground-state charge density and reduce the computational overheads. We develop a grid-centred structural representation, based on Jacobi and Legendre polynomials combined with a linear regression, to accurately learn the converged DFT charge density. This integrates into a ML pipeline that can return any density-dependent observable, including energy and forces, at the quality of a converged DFT calculation, but at a fraction of the computational cost. Fast scanning of energy landscapes and producing starting densities for the DFT self-consistent cycle are among the applications of our scheme. (AU)

• Processo FAPESP: 19/04527-0 - Interface entre isolantes topológicos cristalinos e materiais 2D-trivial: estudo de proximidade via defeitos
• Beneficiário: Bruno Focassio
• Modalidade de apoio: Bolsas no Brasil - Doutorado Direto
• Processo FAPESP: 17/02317-2 - Interfaces em materiais: propriedades eletrônicas, magnéticas, estruturais e de transporte
• Beneficiário: Adalberto Fazzio
• Modalidade de apoio: Auxílio à Pesquisa - Temático
• Processo FAPESP: 21/12204-6 - Aprendizado de máquina para propriedades de materiais bidimensionais
• Beneficiário: Bruno Focassio
• Modalidade de apoio: Bolsas no Exterior - Estágio de Pesquisa - Doutorado Direto