Belief propagation in loopy Bayesian networks: experimental results

We investigate the hypothesis that belief propagation "converges with high probability to the correct decision" on a broad class of loopy belief networks. Experimental results of belief propagation on bipartite graphs, which are the belief networks for turbo codes, are presented.

Saved in:
Bibliographic Details
Published inProceedings of IEEE International Symposium on Information Theory p. 232
Main Authors Horn, G.B., McEliece, R.J.
Format Conference Proceeding
LanguageEnglish
Published IEEE 1997
Subjects
Bayesian methods
Belief propagation
Bipartite graph
Contracts
Decoding
Inference algorithms
Information theory
Intelligent networks
Scholarships
Turbo codes
Online AccessGet full text

Cover

More Information
Summary:We investigate the hypothesis that belief propagation "converges with high probability to the correct decision" on a broad class of loopy belief networks. Experimental results of belief propagation on bipartite graphs, which are the belief networks for turbo codes, are presented.
ISBN:9780780339569
0780339568
DOI:10.1109/ISIT.1997.613147