Title: Verifying the Security of Cyber-Physical Systems
Abstract: Cyber-Physical Systems (CPS) operate in a wide range of applications, where their security is a major concern. Verifying security properties, commonly specified by temporal logics, can be difficult, due to either the lack of accurate models or the models' complexity. In this talk, several recent efforts on addressing this problem will be presented. Presentation includes a new model reduction method for building approximate bi-simulations between complex CPS and finite state Markov chains and a few new statistical model checking techniques using anti-correlated samples and reinforcement learning.
Background: Dr. Yu Wang is currently a postdoctoral associate of Electrical and Computer Engineering at Duke University. He received his Ph.D. degree in Mechanical Engineering from the University of Illinois at Urbana-Champaign. His research interests include the security and privacy in cyber-physical systems. His recent paper was selected one of the best paper finalists of EMSOFT ''19. His Ph.D. dissertation "Statistical Verification and Differential Privacy in Cyber-Physical Systems" was nominated for the CSL Ph.D. Thesis Award of the Coordinated Science Laboratory of UIUC.