Lunch and Learn with Arunesh Sinha
12:00pm - 1:00pm
Engineering & Computer Science, Room 401, Classroom
Join us for a Lunch and Learn with Arunesh Sinha as he speaks on AI and Multiagent Systems for Safety and Security. Food will be available on a first come, first serve basis.
Understanding the complex defender-adversary interaction in any adversarial interaction allows for the design of intelligent and adaptive defense. Game theory is a natural model for such multi-agent interaction. However, significant challenges need to be overcome in order to apply game theory in practice. In this talk, I will present my work on addressing two such challenges: scalability and learning opponent behavior in games. First, I will present a game model of screening of passengers at airports and a novel optimization approach based on randomized allocation and disjunctive programming techniques to solve large instances of the problem. This airport screening work was done in collaboration with the Transport Security Administration in USA. Next, I will present shortcomings of learning adversary behavior and planning optimal defensive actions based on the learned model. A formal learning theory analysis of the learning module reveals why such learning and planning composition fails. I will also present a technique to fix this problem in a security game setting. This emphasizes the need of formal compositional reasoning when using learning as a component in large multi-agent systems.