Title :
Linear Response Formula in Generalized Belief Propagation for Probabilistic Inference
Author :
Tanaka, Kazuyuki
Author_Institution :
Graduate Sch. of Inf. Sci., Tohoku Univ., Sendai
Abstract :
Linear response formulas for the generalized belief propagation in approximate inference are derived by using generalized belief propagation. The linear response formulas can give us the marginal probability between every pair of nodes and we obtain good accuracy in some examples of practical probabilistic inferences
Keywords :
belief maintenance; inference mechanisms; probability; uncertainty handling; approximate inference; generalized belief propagation; linear response formula; probabilistic inference; Belief propagation; Decoding; Digital communication; Digital signal processing; Electronic mail; Information processing; Machine learning; Medical diagnosis; Probability; Random variables;
Conference_Titel :
Computational Intelligence for Modelling, Control and Automation, 2005 and International Conference on Intelligent Agents, Web Technologies and Internet Commerce, International Conference on
Conference_Location :
Vienna
Print_ISBN :
0-7695-2504-0
DOI :
10.1109/CIMCA.2005.1631341
