Description:

The SkAI Institute will host a colloquium featuring Yuanyuan Shi, Assistant Professor in the Department of Electrical and Computer Engineering at the University of California San Diego.

Coffee and pastries will be provided prior to the event. Lunch will be provided afterwards.

Talk title: “Neural Operator Learning for Control”

Abstract: In this talk, we present a unified framework of physics-informed neural operator learning for control of PDE-governed systems. Classical model-based approaches such as PDE backstepping provide strong stability guarantees but remain computationally intensive due to the need to solve kernel PDEs and compute spatially distributed gains. We develop a neural-operator-based approximation framework that learns the mapping from system coefficients to backstepping gains with desired accuracy. Using these approximated gains, we demonstrate that NOC can accelerate PDE feedback synthesis by up to three orders of magnitude while maintaining rigorous stability guarantees. 

Register by Monday, March 16 at 12:00 p.m. at the link below.