Berlin 2024 – wissenschaftliches Programm

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

O: Fachverband Oberflächenphysik

O 82: Electronic Structure Theory I

O 82.3: Vortrag

Donnerstag, 21. März 2024, 11:00–11:15, MA 043

Surrogate models for the electron density and related scalar fields — Joseph Abbott¹, •Raymond Chong¹, Alan Lewis², and Michele Ceriotti¹ — ¹Laboratory of Computational Science and Modeling (COSMO), IMX, École Polytechnique Fédérale de Lausanne, Switzerland — ²University of York, United Kingdom

The electron density is a central quantity in electronic structure calculations and a fundamental property of molecules and materials, allowing access to in principle any ground state electronic property. Density-functional theory (DFT) is an ab initio method that calculates the electron density of a system by solving self- consistently the Kohn-Sham equations. However, the cubic scaling in system size makes calculations for large systems intractable. As a complementary approach, machine learning (ML) surrogate models are being developed that directly predict the self-consistent electron density at a fraction of the cost of DFT, and with more favourable linear scaling. One framework in particular focuses on learning the coefficients of basis functions fitted to the real-space density using a local equivariant ML model.

An end-to-end pipeline for the training and prediction of the the electron density is presented here. Built on top of a modular and scalable software stack, the 'bypassing' of the Kohn-Sham equations to access the density of larger systems becomes increasingly possible. Furthermore, the framework can be applied to scalar field related to the electron density, opening the door to interesting applications such as the ML-driven imaging of scanning tunneling microscopy (STM).

Keywords: machine-learning; local-density-of-states; electron-density