Departmental Papers (CIS)
Date of this Version
7-2011
Document Type
Conference Paper
Recommended Citation
Kareem Amin, Michael J. Kearns, and Umar Syed, "Graphical Models for Bandit Problems", . July 2011.
Abstract
We introduce a rich class of graphical models for multi-armed bandit problems that permit both the state or context space and the action space to be very large, yet succinctly specify the payoffs for any context-action pair. Our main result is an algorithm for such models whose regret is bounded by the number of parameters and whose running time depends only on the treewidth of the graph substructure induced by the action space.
Date Posted: 24 July 2012
Comments
Amin, K., Kearns, M., & Syed, U., Graphical Models for Bandit Problems, 27th Conference on Uncertainty in Artificial Intelligence, July 2011, http://arxiv.org/abs/1202.3782