Research Interests

artificial intelligence, machine learning, game theory, multi-agent systems

Brief Research Statement

As of Spring 2009, I am working under the advisement of Dr. Dana Nau. My research is currently focused on identifying local pathology in subtrees of potentially non-pathological game trees. Pathology in game tree search is a phenomenon discovered nearly 30 years ago by Dana Nau in which deeper search may actually result in poorer decision quality.

Publications

[1] Brandon Wilson, Inon Zuckerman, and Dana Nau, "Modeling Social Preferences in Multi-player Games," The International Conference on Autonomous Agents and Multiagent Systems, Taipei, Taiwan, 2011. [pdf] [slides]

[2] Brandon Wilson, "Improving Game-tree Search by Incorporating Error Propagation and Social Orientations (Extended Abstract)," The International Conference on Autonomous Agents and Multiagent Systems, Taipei, Taiwan, 2011. [pdf]

[3] Brandon Wilson, Austin Parker, and Dana Nau, "Error Minimizing Minimax: Avoiding Search Pathology in Game Trees," The International Symposium on Combinatorial Search, Lake Arrowhead, California, 2009. [pdf] [slides]

[4] Brandon Wilson. "Test-cost Sensitive Regression for Planner Runtime Prediction." Master's Thesis, University of Maryland Baltimore County. 2008. [pdf]

[5] Mark Roberts, Adele Howe, Brandon Wilson and Marie desJardins, "What makes planners predictable?," Proceedings of the Eighteenth International Conference on Automated Planning and Scheduling, 2008. [pdf]

[6] Brandon Wilson and Marie desJardins, "Forming Stable, Overlapping Coalitions in an Open Mulit-agent System." Working Notes of the AAAI Fall Symposium on Regarding the Intelligence in Distributed Intelligent Systems, 2007.