MC 5501
Peter Yichen Chen, Massachusetts Institute of Technology
Title
Neural PDE: Towards AI-enhanced Physics Simulation
Abstract
Physics simulation has become the third pillar of science and engineering, along with theory and experiments. The overarching objective of my cross-disciplinary research is to democratize physics simulation. This is achieved through a thoughtful fusion of cutting-edge AI methodologies and classical numerical methods. In this talk, I will introduce two research threads that align with this overarching theme. These threads will harness various machine learning tools (e.g., neural fields) to advance physics simulations’ accuracy and speed. A recurring theme in all two threads is the exceptional generalization capabilities of these ML-enhanced simulations, thanks to the careful incorporation of partial differential equations (PDEs) as an inductive bias.