DÍA / HORA: Martes 7 de abril del 2026 / 16:30 – 17:10
EXPOSITOR: Elie Bretin, INSA Lyon Institut Camille Jordan ICJ UMR CNRS 5208
RESUMEN: In this talk, we will focus on the use of neural networks as tools for numerically approximating the solutions of phase field models. These methods are particularly effective for approximating evolving interfaces, such as the mean curvature flow and its famous Allen-Cahn equation. We will explain how to build neural network structures strongly inspired by the numerical schemes of the Allen-Cahn equation in order to exploit them in more complex situations. We are interested, for example, in the case of crystal growth models, where anisotropic diffusion operators increase certain numerical stabilities, or in the case of non-oriented interfaces, for which standard approaches are not suitable. This work was carried out in collaboration with R. Denis, S. Masnou and G. Terii.
IDIOMA: Inglés
LUGAR: Sala de Seminarios 5to Piso, Departamento de Ingeniería Matemática, FCFM, Universidad de Chile.
DIRECCIÓN: Beauchef 851, Torre Norte, Piso 5, Departamento de Ingeniería Matemática, FCFM, Universidad de Chile.
MODALIDAD: Presencial y transmisión zoom (link zoom seminario online)