Home Page

Papers

Submissions

News

Editorial Board

Open Source Software

Proceedings (PMLR)

Transactions (TMLR)

Search

Statistics

Login

Frequently Asked Questions

Contact Us



RSS Feed

Improving Variational Methods via Pairwise Linear Response Identities

Jack Raymond, Federico Ricci-Tersenghi; 18(6):1−36, 2017.

Abstract

Inference methods are often formulated as variational approximations: these approxima- tions allow easy evaluation of statistics by marginalization or linear response, but these estimates can be inconsistent. We show that by introducing constraints on covariance, one can ensure consistency of linear response with the variational parameters, and in so doing inference of marginal probability distributions is improved. For the Bethe approximation and its generalizations, improvements are achieved with simple choices of the constraints. The approximations are presented as variational frameworks; iterative procedures related to message passing are provided for finding the minima.

[abs][pdf][bib]       
© JMLR 2017. (edit, beta)