Computer Science Seminar Series
Most Relevant Explanation in Bayesian Networks
3:00 p.m. Wednesday, March 25, 2009
Weir Hall, Room 235
Changhe Yuan
Director of Uncertainty Reasoning Laboratory
Assistant Professor of Computer Science and Engineering
Mississippi State University
Abstract:
Many existing explanation methods in Bayesian networks, such as Maximum a Posteriori (MAP) assignment and Most Probable Explanation (MPE), generate complete assignments for target variables. A priori, the set of target variables is often large, but only a few of them may be most relevant in explaining given evidence. It is desirable to generate more concise explanations. In this talk, I will discuss a new framework called Most Relevant Explanation (MRE) for finding explanations consisting of the most relevant target variables. I will also discuss an approximate algorithm for solving MRE based on Reversible Jump MCMC and simulated annealing. Both theoretical analysis and empirical results show that the new approach is a promising approach for finding explanations in Bayesian networks.
Biography:
Dr. Changhe Yuan is an Assistant Professor of Computer Science and Engineering at Mississippi State University and director of the Uncertainty Reasoning Laboratory (URL Lab). His research lab focuses on developing algorithms for probabilistic modeling and reasoning and tackling real-world problems in interdisciplinary areas, such as diagnosis/prognosis, risk management, computational biology, and education. Dr. Yuan received his PhD from the Intelligent Systems Program at University of Pittsburgh in 2006.