Friday, July 15th, 2022. Time: 12 noon

Hybrid! Place: Room C2-212, C2 Building, UPC Campus Nord, Barcelona | Link of the session: https://meet.google.com/ekd-qqgp-rec

See video on Youtube

ABSTRACT

In this talk, a new numerical framework, employing physics-informed neural networks, to find a smooth self-similar solution for the Boussinesq equations will be explained. The solution in addition corresponds to an asymptotic self-similar profile for the 3-dimensional Euler equations in the presence of a cylindrical boundary and represents a precise description of a possible blow-up scenario. The new numerical framework is shown to be both robust and readily adaptable to other equations. Quanta Magazine wrote a piece: https://www.quantamagazine.org/deep-learning-poised-to-blow-up-famed-fluid-equations-20220412/ ('Deep Learning Poised to ‘Blow Up’ Famed Fluid Equations') on it recently. Joint work with Yongji Wang, Ching-Yao Lai and Tristan Buckmaster.

SPEAKER CV

He is a Visiting Professor at Brown University and a Distinguished Researcher at University of Barcelona. Before that, he was an Instructor and an Assistant Professor at Princeton. He co-organize the Brown PDE Seminar. His research interests are on the boundary between analysis, partial differential equations, fluid mechanics, spectral geometry, numerical computation and rigorous computer-assisted proofs. His research is partially supported by the European Research Council through ERC Starting Grant 852741. 

See Seminar Programme 2022