October 25, 2024, Webb Hall 1100

Control and machine learning are two high-impact fields. The development of next-generation intelligent systems, such as self-driving vehicles, humanoid robots, smart buildings, and automated healthcare, requires a rapprochement between these two areas. This talk will cover some fundamental connections between learning and control and discuss how generative AI is changing our perspectives on how these two fields are bridged. The first half of this talk focuses on some connections of learning and control that were built before the era of generative AI. We will discuss how to tailor control-theoretic tools, such as quadratic constraints, to unify the development of learning algorithms and models. Additionally, we will borrow tools from reinforcement learning, non-convex optimization, and neural network verification to push the boundaries of robust control. In the second half, we will focus on how generative AI shifts the research on the interplay between control and machine learning. We will present our recent work exploring the capabilities of state-of-the-art large language models (LLMs) in solving control design problems. If time permits, we will also discuss two additional papers that rethink the concept of controllability in the context of diffusion models and LLM jailbreaking.

Bin Hu received the B.Sc. in Theoretical and Applied Mechanics from the University of Science and Technology of China in 2008, and received the M.S. in Computational Mechanics from Carnegie Mellon University in 2010. He received the Ph.D. in Aerospace Engineering and Mechanics at the University of Minnesota in 2016. Between July 2016 and July 2018, he was a postdoctoral researcher in the Wisconsin Institute for Discovery at the University of Wisconsin-Madison. He is currently an assistant professor in the Department of Electrical and Computer Engineering at the University of Illinois Urbana-Champaign and affiliated with the Coordinated Science Laboratory. His research focuses on building fundamental connections between control theory and machine learning. He received the NSF CAREER Award and the Amazon Research Award in 2021, and the O. Hugo Schuck Best Paper Award in 2024 for his joint work "Learning the Kalman Filter with Fine-grained Sample Complexity" with Xiangyuan Zhang and Tamer Başar.