Campus Event Calendar
Event Entry
What and Who
Message Passing Algorithms
Megha Khosla
IMPRS-CS
Master Seminar
IMPRS Master student
AG 1, AG 3, AG 4, AG 5, SWS, RG1, MMCI
Public Audience
English
Note: We use this to send email in the morning.
Date, Time and Location
Wednesday, 17 June 2009
13:05
60 Minutes
E1 4
024
Saarbrücken
Abstract
Important inference problems in computer vision, error-correcting coding theory, and artificial intelligence can all be formulated as the computation of marginal probabilities on factor graphs. We focus on the Belief Propagation (BP) Algorithm, which is an efficient way for solving inference problems based on passing local messages.
Belief Propagation Algorithm is exact only when the factor graph is a tree. Convergence and accuracy cannot be guaranteed for a general graph. We propose a method for improving BP that takes into account the influence of cycles in the factor graph. We prove its convergence and exactness for a single cycle and are working towards generalizing the approach.
Belief Propagation Algorithm is exact only when the factor graph is a tree. Convergence and accuracy cannot be guaranteed for a general graph. We propose a method for improving BP that takes into account the influence of cycles in the factor graph. We prove its convergence and exactness for a single cycle and are working towards generalizing the approach.
Contact
imprs
0681 93 25 225
--email hidden
passcode not visible
logged in users only
Stephanie Jörg, 06/16/2009 09:09 -- Created document.