ECE Seminar by Ming Jin: Trustworthy Reinforcement Learning for Safety-Critical Systems

Armour College of Engineering’s Department of Electrical and Computer Engineering will welcome Ming Jin, an Assistant Professor in the Bradley Department of Electrical and Computer Engineering at Virginia Tech to present a lecture, “Trustworthy Reinforcement Learning for Safety-Critical Systems.” online Friday, November 12 from 12:45-1:45 p.m.

Use this link for more information and to sign up to attend: https://www.iit.edu/events/ece-seminar-ming-jin-trustworthy-reinforcement-learning-safety-critical-systems

Abstract: Assurance is integral to trust and conducive to widespread adoption in energy infrastructures. However, such safety and performance guarantees are currently lacking for learning-based control methods, especially when high-capacity models such as deep neural networks are involved. In this talk, I will discuss some recent works towards addressing this challenge. In the first part, I will discuss about a control-theoretic framework for safe reinforcement learning, with applications to the control of partially observable, nonlinear, and uncertain systems using (deep) neural networks. In the second part, I will present an implicit reinforcement learning framework that leverages the synergistic strength of optimization and reinforcement learning for online adaptivity and trustworthy sequential decision making. Throughout the talk, I will use power systems as the main examples for illustration. The papers that I will discuss are included as follows.

  1. Zeroth-Order Implicit Reinforcement Learning for Sequential Decision Making in Distributed Control Systems
  2. Imitation Learning with Stability and Safety Guarantees
  3. Stability-Certified Reinforcement Learning: A Control-Theoretic Perspective
  4. Recurrent Neural Network Controllers Synthesis with Stability Guarantees for Partially Observed Systems

Biography: Dr. Ming Jin is an Assistant Professor in the Bradley Department of Electrical and Computer Engineering at Virginia Tech. He received the B.Eng. degree with honor from Hong Kong University of Science and Technology in 2012, the Ph.D. degree from the Department of Electrical Engineering and Computer Sciences at University of California, Berkeley in 2017, and was a postdoctoral scholar at the Department of Industrial Engineering and Operations Research at University of California, Berkeley from 2018 to 2020 before joining Virginia Tech. Jin’s work has been recognized by three best paper awards and featured in multiple media outlets, including IEEE Spectrum, Berkeley Engineer Magazine, and MIT Technology Review. His research interests and expertise include trustworthy AI, control theory, optimization, machine learning, cyber-physical systems, and power systems.

REGISTER/ZOOM LINK