Presentation 2003/12/1
Accuracy of the MCMC Methods in Learning Models with Singularities
Katsuyuki TAKAHASHl, Sumio WATANABE,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In the Bayesian learning, the stochastic complexity has been used for model selection arid optimization of hyper parameters. However, since there has been no method to calculate analytically the stochastic complexity, we could not evaluate how precise the MCMC method is. In this paper, we apply the algebraic geometrical method to the stochastic complexity of singular learning machines, and clarify the properties of the MCMC method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) stochastic complexity / MCMC method / regular models / learning models with singularity / algebraic
Paper # NC2003-103
Date of Issue

Conference Information
Committee NC
Conference Date 2003/12/1(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Vice Chair

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Accuracy of the MCMC Methods in Learning Models with Singularities
Sub Title (in English)
Keyword(1) stochastic complexity
Keyword(2) MCMC method
Keyword(3) regular models
Keyword(4) learning models with singularity
Keyword(5) algebraic
1st Author's Name Katsuyuki TAKAHASHl
1st Author's Affiliation Department of Advanced Applied Electronics Tokyo Institute of Technology()
2nd Author's Name Sumio WATANABE
2nd Author's Affiliation P&I Lab.Tokyo Institute
Date 2003/12/1
Paper # NC2003-103
Volume (vol) vol.103
Number (no) 490
Page pp.pp.-
#Pages 6
Date of Issue