Last week, we had the ERCOFTAC ML4Fluids conference at the Centrum Wiskunde & Informatica in Amsterdam. I was happy to contribute to the conference organitzation together with the Benjamin Sanderse, Richard Dwight, Nguyen Anh Khoa Doan and Wouter Edeling.
From my group, we had the presentations from PhD students Manuel Cabral on hard-constraints for physics-based neural networks, and Eric Chillón on acceleration of algebraic multigrid linear solvers using graph neural networks. My collaborator Hussam Alhussein (sababha) presented the work of reinforcement learning for vortex-induced vibrations in experiments.
It was great to experience the conference from the organizational size, and I was also inspired by the many quality talks we had. Also the keynotes were a highlight of the conference, including topics such as information-theoretic causality (Adrian Lozano Duran), solution blow-up of Euler equations through multistage neural networks (Ching-Yao Lai), ML-based weather forecast (Daniel Worrall), and AI for fluid dynamics (Gianluca Iaccarino).
Find all the abstracts and keynote slides in the ML4Fluids website.
Thanks to everyone who attended and made the conference possible!
