Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

MM: Fachverband Metall- und Materialphysik

MM 43: Data Driven Material Science: Big Data and Workflows V

MM 43.1: Vortrag

Mittwoch, 20. März 2024, 15:45–16:00, C 243

The MALA Package - Transferable and Scalable Electronic Structure Simulations Powered by Machine Learning — •Lenz Fiedler and Attila Cangi — Helmholtz-Zentrum Dresden-Rossendorf / CASUS

Interactions between electrons and nuclei determine all materials properties, and modeling these interactions is of paramount importance to pressing scientific questions. However, even with the most advanced electronic structure tools, such as density functional theory (DFT), electronic structure simulations are usually restricted to a few thousand atoms.

Machine-learning DFT (ML-DFT) tackles this challenge by providing rapid access to observables of interest. Most current ML-DFT methodologies focus on the mapping between ionic configurations and scalar observables, rather than a full prediction of electronic structure. The Materials Learning Algorithms (MALA) python package addresses this gap, by providing a user-friendly framework to construct ML-DFT models that give access to a range of electronic structure observables, such as the electronic density, DOS and total free energy. The usefulness of these models across phase boundaries [1], multiple length scales [2] and temperature ranges [3] has been amply demonstrated. In this talk, a general introduction to the framework along with its computational properties and capabilities is given.

[1] J. A. Ellis et. al, Phys. Rev. B, 2021, 104, 035120 ** [2] L. Fiedler et. al, npj Comput. Mater., 2022, 9, 115

[3] L. Fiedler et. al, Phys. Rev. B, 2023, 108, 125146

Keywords: Density Functional Theory; Machine Learning; Scientific Software; Surrogate Model