Novel iteration schemes for the Cluster Variation Method

Part of Advances in Neural Information Processing Systems 14 (NIPS 2001)

Bibtex Metadata Paper

Authors

Hilbert Kappen, Wim Wiegerinck

Abstract

The Cluster Variation method is a class of approximation meth(cid:173) ods containing the Bethe and Kikuchi approximations as special cases. We derive two novel iteration schemes for the Cluster Vari(cid:173) ation Method. One is a fixed point iteration scheme which gives a significant improvement over loopy BP, mean field and TAP meth(cid:173) ods on directed graphical models. The other is a gradient based method, that is guaranteed to converge and is shown to give useful results on random graphs with mild frustration. We conclude that the methods are of significant practical value for large inference problems.