A whirlwind tour of the SINDy ecosystem

Abstract: Identifying models from sparse data is a central challenge in mathematical physics, with a rich history of developments in fluid dynamics. The form of the dynamics is typically either constrained via prior knowledge, as in Galerkin projection, or a particular models structure is chosen heuristically and parameters are optimized to match the data. Simultaneous identiification of the model structure and parameters from data is considerably more challenging as there are combinatorially many possible such structures. Over the past decade, the Sparse Identification of Nonlinear Dynamics (SINDy) methodology has gained a lot of traction for such modeling problems.

In this talk, we will give a whirlwind tour of the SINDy ecosystem. Starting with the identification of simple nonlinear ordinary differential equations, we will illustrate how the basic SINDy methodology can be extended to enforce known symmetries or guaranteed stability, identify dynamics defined by rational functions or stochastic differential equations, and extend it with uncertainty quantification capabilities. All the examples presented during the talk make use of the pySINDy package, an open-source python package providing state-of-the-art implementations of the various SINDy flavours.

‼️ Le séminaire peut être suivi en ligne via le lien : https://univ-amu-fr.zoom.us/j/95011527683?pwd=bTeNvaaZY75buaa7bRL0pcucXSASm7.1

Le mardi 2 juin 2026 à 11h00 / Amphithéâtre François Canac, LMA

Voir en ligne : plus d’informations concernant l’orateur