Dresden 2020 – scientific programme

The DPG Spring Meeting in Dresden had to be cancelled! Read more ...

Parts | Days | Selection | Search | Updates | Downloads | Help

MM: Fachverband Metall- und Materialphysik

MM 30: Poster Session II

MM 30.34: Poster

Tuesday, March 17, 2020, 18:15–20:00, P4

Neural network for learning and predicting tight-binding parameters — •Till Hanke, Jürgen Henk, and Ingrid Mertig — Martin-Luther-Universität Halle-Wittenberg

Tight-binding approaches have two major advantages: they allow for an intuitive interpretation of electronic structures and to perform large-scale electronic-structure calculations. However, parameter sets, either DFT-based or empirical, are often available only for simple bulk systems.

We report on artificial neural networks which can predict Slater-Koster tight-binding parameters for heterogeneous systems. The networks are trained using parameter sets for elemental materials. These sets will be used for electronic-structure and transport calculations (on the femtosecond timescale) for which accurate descriptions of complex interfaces are essential.