File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
  • Find via Find It@HKUL
Supplementary

Article: Convergence analysis of distributed inference with vector-valued Gaussian belief propagation

TitleConvergence analysis of distributed inference with vector-valued Gaussian belief propagation
Authors
Issue Date2017
PublisherJournal of Machine Learning Research. The Journal's web site is located at http://mitpress.mit.edu/jmlr
Citation
Journal of Machine Learning Research, 2017, v. 18 n. 1, p. 6302-6339 How to Cite?
AbstractThis paper considers inference over distributed linear Gaussian models using factor graphs and Gaussian belief propagation (BP). The distributed inference algorithm involves only local computation of the information matrix and of the mean vector, and message passing between neighbors. Under broad conditions, it is shown that the message information matrix converges to a unique positive definite limit matrix for arbitrary positive semidefinite initialization, and it approaches an arbitrarily small neighborhood of this limit matrix at an exponential rate. A necessary and sufficient convergence condition for the belief mean vector to converge to the optimal centralized estimator is provided under the assumption that the message information matrix is initialized as a positive semidefinite matrix. Further, it is shown that Gaussian BP always converges when the underlying factor graph is given by the union of a forest and a single loop. The proposed convergence condition in the setup of distributed linear Gaussian models is shown to be strictly weaker than other existing convergence conditions and requirements, including the Gaussian Markov random field based walk-summability condition, and applicable to a large class of scenarios.
Persistent Identifierhttp://hdl.handle.net/10722/259296
ISSN
2015 Impact Factor: 2.45
2015 SCImago Journal Rankings: 1.648

 

DC FieldValueLanguage
dc.contributor.authorDu, J-
dc.contributor.authorMa, S-
dc.contributor.authorWu, YC-
dc.contributor.authorKar, S-
dc.contributor.authorMoura, J-
dc.date.accessioned2018-09-03T04:04:42Z-
dc.date.available2018-09-03T04:04:42Z-
dc.date.issued2017-
dc.identifier.citationJournal of Machine Learning Research, 2017, v. 18 n. 1, p. 6302-6339-
dc.identifier.issn1532-4435-
dc.identifier.urihttp://hdl.handle.net/10722/259296-
dc.description.abstractThis paper considers inference over distributed linear Gaussian models using factor graphs and Gaussian belief propagation (BP). The distributed inference algorithm involves only local computation of the information matrix and of the mean vector, and message passing between neighbors. Under broad conditions, it is shown that the message information matrix converges to a unique positive definite limit matrix for arbitrary positive semidefinite initialization, and it approaches an arbitrarily small neighborhood of this limit matrix at an exponential rate. A necessary and sufficient convergence condition for the belief mean vector to converge to the optimal centralized estimator is provided under the assumption that the message information matrix is initialized as a positive semidefinite matrix. Further, it is shown that Gaussian BP always converges when the underlying factor graph is given by the union of a forest and a single loop. The proposed convergence condition in the setup of distributed linear Gaussian models is shown to be strictly weaker than other existing convergence conditions and requirements, including the Gaussian Markov random field based walk-summability condition, and applicable to a large class of scenarios.-
dc.languageeng-
dc.publisherJournal of Machine Learning Research. The Journal's web site is located at http://mitpress.mit.edu/jmlr-
dc.relation.ispartofJournal of Machine Learning Research-
dc.titleConvergence analysis of distributed inference with vector-valued Gaussian belief propagation-
dc.typeArticle-
dc.identifier.emailWu, YC: ycwu@eee.hku.hk-
dc.identifier.authorityWu, YC=rp00195-
dc.identifier.hkuros289199-
dc.identifier.volume18-
dc.identifier.issue1-
dc.identifier.spage6302-
dc.identifier.epage6339-
dc.publisher.placeUnited States-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats