Proceedings of the Thirty-Second Conference on Uncertainty in Artificial Intelligence, UAI 2016, June 25-29, 2016, New York City, NY, USA
نویسندگان
چکیده
Stackelberg games are two-stage games in whichthe first player (called the leader) commits to astrategy, after which the other player (the fol-lower) selects a best-response. These types ofgames have seen numerous practical applicationin security settings, where the leader (in thiscase, a defender) must allocate resources to pro-tect various targets. Real world applications in-clude the scheduling of US federal air marshalsto international flights, and resource allocation atLAX airport. However, the best known algorithmfor solving general Stackelberg games requiressolving Integer Programs, and fails to scale be-yond a few (significantly smaller than 100) num-ber of leader actions, or follower types. In thispaper, we present a new gradient-based approachfor solving large Stackelberg games in securitysettings. Large-scale control problems are oftensolved by restricting the controller to a rich pa-rameterized class of policies; the optimal controlcan then be computed using Monte Carlo gradi-ent methods. We demonstrate that the same ap-proach can be taken in a strategic setting. Weevaluate our approach empirically, demonstrat-ing that it can have negligible regret against theleader’s true equilibrium strategy, while scalingto large games.
منابع مشابه
Needed: Global Collaboration for Comparative Research on Cities and Health
Over half of the world’s population lives in cities and United Nations (UN) demographers project an increase of 2.5 billion more urban dwellers by 2050. Yet there is too little systematic comparative research on the practice of urban health policy and management (HPAM), particularly in the megacities of middle-income and developing nations. We make a case for creating a global database on citie...
متن کاملMeeting the Challenge of Diabetes in China
China’s estimated 114 million people with diabetes pose a massive challenge for China’s health policy-makers who have significantly extended health insurance coverage over the past decade. What China is doing now, what it has achieved, and what remains to be done should be of interest to health policy-makers, worldwide. We identify the challenges posed by China’s two pr...
متن کاملUAI '00: Proceedings of the 16th Conference in Uncertainty in Artificial Intelligence, Stanford University, Stanford, California, USA, June 30 - July 3, 2000
Let's read! We will often find out this sentence everywhere. When still being a kid, mom used to order us to always read, so did the teacher. Some books are fully read in a week and we need the obligation to support reading. What about now? Do you still love reading? Is reading only for you who have obligation? Absolutely not! We here offer you a new book enPDFd uai 00 proceedings of the 16th c...
متن کاملProceedings of the Thirty-Third Conference on Uncertainty in Artificial Intelligence, UAI 2017, Sydney, Australia, August 11-15, 2017
متن کامل
UAI '01: Proceedings of the 17th Conference in Uncertainty in Artificial Intelligence, University of Washington, Seattle, Washington, USA, August 2-5, 2001
By reading, you can know the knowledge and things more, not only about what you get from people to people. Book will be more trusted. As this uai 01 proceedings of the 17th conference in uncertainty in artificial intelligence, it will really give you the good idea to be successful. It is not only for you to be success in certain life you can be successful in everything. The success can be start...
متن کامل