TY - JOUR AU - Perrault, Andrew AU - Wilder, Bryan AU - Ewing, Eric AU - Mate, Aditya AU - Dilkina, Bistra AU - Tambe, Milind PY - 2020/04/03 Y2 - 2024/03/28 TI - End-to-End Game-Focused Learning of Adversary Behavior in Security Games JF - Proceedings of the AAAI Conference on Artificial Intelligence JA - AAAI VL - 34 IS - 02 SE - AAAI Technical Track: Computational Sustainability DO - 10.1609/aaai.v34i02.5494 UR - https://ojs.aaai.org/index.php/AAAI/article/view/5494 SP - 1378-1386 AB - <p>Stackelberg security games are a critical tool for maximizing the utility of limited defense resources to protect important targets from an intelligent adversary. Motivated by green security, where the defender may only observe an adversary's response to defense on a limited set of targets, we study the problem of learning a defense that <em>generalizes</em> well to a new set of targets with novel feature values and combinations. Traditionally, this problem has been addressed via a <em>two-stage</em> approach where an adversary model is trained to maximize predictive accuracy without considering the defender's optimization problem. We develop an end-to-end <em>game-focused</em> approach, where the adversary model is trained to maximize a surrogate for the defender's expected utility. We show both in theory and experimental results that our game-focused approach achieves higher defender expected utility than the two-stage alternative when there is limited data.</p> ER -