LiPhy Grenoble
1 February 2021: One event
-
Invited Talks
En savoir plus : Invited TalksMonday 1 February 14:00-15:30 - Remy Kusters - Physics Inspired Machine Learning Lab, CRI, Paris Deep learning driven model discovery for physics Résumé : With the growing need for interpretability of machine learning-based predictions in quantitative science our team develops DeepMoD: An open source Deep learning based framework for Model Discovery. Model discovery aims at finding interpretive models in the form of partial differential equations from large spatio-temporal data-sets. By integrating advances in symbolic and statistical machine learning DeepMoD seeks an interpretable model from noisy and oftentimes very sparse data-sets. We illustrate this approach on several problems in the context of (bio)physics, mechanics and fluid dynamics, such as the Burgers’, Korteweg-de Vries, advection-diffusion and Keller-Segel equations.
contact: S. RafaiLieu : online