As part of the PhD course "Intelligent Decision Support Systems", Daniele Codecasa, PhD student from University of Milano Bicocca, will give the lecture
"Exact Inference in Bayesian Networks & Influence Diagrams"
Tue 20/09/2011, 2pm, Seminar room, Computer Science Department
Bayesian networks are a framework used to represent probability distribution over a set of variables. This distribution is represented using conditional probability in order to structurize the problem and simplify the probabilities representation. Bayesian networks are widely used to model real contexts. In this seminar is showed the lazy propagation algorithm over junction trees. This algorithm is used to make exact inference over bayesian networks. There will be a brief introduction of influence diagrams, a framework used to make decision under uncertainty. Finally, the application of the previous algorithm to make inference in order to take decisions is explained.
Speaker: Daniele Codecasa Ph.D. Student
Department of Informatics, Systems and Communication
Building U14, room T004 viale Sarca, 336 University of Milano-Bicocca 20126 Milano, Italy
Tel.: +39 02 6448 7852
Web page: http://www.mad.disco.unimib.it/doku.php/people/daniele_codecasa