Adversary Bounded Rationality in Green Security Domains: Handling Payoff Uncertainty and Elicitation

Than Nguyen
Wednesday, April 1, 2015 - 14:00
PHE 333
Research on Stackelberg Security Games (SSG) has recently shifted to green security domains, e.g., protecting wildlife from illegal poaching. Previous research on this topic has advocated the use of behavioral (bounded rationality) models of adversaries in SSG. We, for the first time, provides validation of these behavioral models based on real-world data from a wildlife park. We next introduce the first algorithm to handle payoff uncertainty – an important concern in green security domains – in the presence of such adversary behavioral models. Finally, given the availability of mobile sensors such as Unmanned Aerial Vehicles in green security domains, we introduce new payoff elicitation strategies to strategically reduce uncertainty over multiple targets at a time.

Bio: Thanh Nguyen is a PhD student in the Dept. of Computer Science. She is working with Prof. Milind Tambe at the Teamcore research group. Her research interests include game-theoretic modeling, robust solution techniques, and behavioral modeling for real-world security applications. In particular, she has been focusing on developing efficient robust algorithms for handling worst-case cenario of uncertainties in Stackelberg Security Games. Furthermore, she is working on predicting decision-making of human adversaries based on their activity history and incorporate that model into computing the optimal protection strategy for security agencies or the defender.