Presentation	2007-03-14 On the Two Kullback Divergences in Approximating Bayesian Posterior Distributions Kazuho WATANABE, Sumio WATANABE,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	Some methods have been proposed and used for approximating Bayesian learning. Although they have provided efficient learning algorithms in various applications, their properties have little been investigated. In this paper, we focus on the two approximation schemes where the Kullback and reversed Kullback information between the approximating distribution and the exact Bayesian posterior distribution are minimized respectively over the factorizable distributions. Considering an example of Bayesian learning in the linear neural network, we show the differences between the two approximating distributions.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	Approximate Bayesian Inference / Kullback Information / Linear Neural Networks
Paper #	NC2006-134
Date of Issue

Conference Information
Committee	NC
Conference Date	2007/3/7(1days)
Place (in Japanese)	(See Japanese page)
Place (in English)
Topics (in Japanese)	(See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To	Neurocomputing (NC)
Language	ENG
Title (in Japanese)	(See Japanese page)
Sub Title (in Japanese)	(See Japanese page)
Title (in English)	On the Two Kullback Divergences in Approximating Bayesian Posterior Distributions
Sub Title (in English)
Keyword(1)	Approximate Bayesian Inference
Keyword(2)	Kullback Information
Keyword(3)	Linear Neural Networks
1st Author's Name	Kazuho WATANABE
1st Author's Affiliation	Department of Computational Intelligence and Systems Science, Tokyo Institute of Technology()
2nd Author's Name	Sumio WATANABE
2nd Author's Affiliation	P&I Lab, Tokyo Institute of Technology
Date	2007-03-14
Paper #	NC2006-134
Volume (vol)	vol.106
Number (no)	588
Page	pp.pp.-
#Pages	6
Date of Issue