Structure-preserving artificial neural networks for moment closures

"Structure-preserving artificial neural networks for moment closures", organizado polo CITMAga. Será impartido por Martin Frank (Steinbuch Centre for Computing- Karlsruhe Institute of Technology)

Data: xoves 27 de abril 

Hora: 10:00 h.

Duración: 1 hora

Lugar: Aula Magna da Facultade de Matemáticas, ou ben en liña a través da ligazón Teams meeting. Conferenciante por Teams.

Abstract:

We present how artificial neural networks can be used to substitute an otherwise costly moment closure for kinetic equations. Specifically, we study entropy closures which have many desirable properties (among them guaranteed hyperbolicity, positivity and entropy decay). We will need to construct very specific artificial neural networks that are guaranteed to preserve these advantages. Furthermore, we address the question of data sampling. Several numerical results are shown.