•   When: Friday, March 11, 2022 from 02:00 PM to 03:00 PM
•   Speakers: Ioannis Karamouzas
•   Location: ZOOM only
•   Export to iCal

Abstract: Given the advancements in artificial intelligence over the last decade along with the significant decrease in hardware cost, autonomous agents are increasingly becoming a part of our lives. For example, household robots such as Roombas are nowadays vacuuming our floors, while immersive game worlds, training simulators, and online communities are commonly populated by animated humans. In this talk, I will highlight our recent efforts to push forward the state-of-the-art in motion planning, with the goal of advancing the AI capabilities of autonomous agents in both virtual and physical environments. First, I will discuss how we can leverage generative models and reinforcement learning with and without the use of expert data for natural control of physically simulated characters in virtual worlds and interactive graphics applications. Second, I will show how data-driven and optimization techniques can be used to derive accurate and robust models of crowd motion and behavior. I will conclude by discussing how the aforementioned techniques can be extended to develop new algorithms for autonomous robot navigation, enabling mobile robots to behave in a more adept manner and to do so with more of the sophistication and intelligence of complex beings. 

 Bio: Ioannis Karamouzas is an assistant professor in the School of Computing at Clemson University. Prior to joining Clemson, he was a postdoctoral associate at the University of Minnesota and received his Ph.D. from Utrecht University. His research focuses on developing motion planning algorithms for autonomous agents, robots and crowds of virtual characters. His work has been integrated into commercial applications  including computer games and pedestrian simulation suites. He is a recipient of an NSF CAREER award, and is currently serving on the editorial board of IEEE Robotics and Automation Letters and Elsevier’s Robotics and Autonomous Systems.