Séminaire de Mécanique d'Orsay

Le Jeudi 11 juillet 2019 à 14h00 - Salle de conférences du LIMSI

Unsupervised identification of dominant balance regimes

Jared Callaham
Department of Applied Mathematics University of Washington

Accurate and interpretable models are critical for understanding and interacting with complex systems. Although highly reliable, detailed physics models are themselves often complicated enough to defy straightforward analysis. A historically successful approach has been to identify limiting regimes in which the dynamics are determined by a dominant balance between terms in the model. We generalize this intuition by using recent model discovery and machine learning methods to search for distinct dynamical regimes in such systems.

Accès Salle de conférences du LIMSI