Presentation 2008-12-09
Evaluation of annealing schadule for PLSA language model adaptaion
Masaharu KATO, Tetsuo KOSAKA, Akinori ITO, Shozo MAKINO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Probabilistic Latent Semantic Analysis (PLSA) is a powerful statistical laguage model. However the PLSA has the local maxima problem. To overcame this problem, the EM annealing algorithm has been proposed. In this paper, we designed annealing schedule β with some continuous functions. As a result, we found that increasing functions and square root functions are the best for annealing schedule. In the experiment, we obtain 28.7% perplexity reduction and 5.3% word error rate reduction.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) probabilistic latent semantic analysis / annealing / EM algorithm / language model adaptation
Paper # NLC2008-32,SP2008-87
Date of Issue

Conference Information
Committee NLC
Conference Date 2008/12/2(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 Natural Language Understanding and Models of Communication (NLC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Evaluation of annealing schadule for PLSA language model adaptaion
Sub Title (in English)
Keyword(1) probabilistic latent semantic analysis
Keyword(2) annealing
Keyword(3) EM algorithm
Keyword(4) language model adaptation
1st Author's Name Masaharu KATO
1st Author's Affiliation Graduate School of Science and Engineering, Yamagata University:Graduate School of Engineering, Tohoku University()
2nd Author's Name Tetsuo KOSAKA
2nd Author's Affiliation Graduate School of Science and Engineering, Yamagata University
3rd Author's Name Akinori ITO
3rd Author's Affiliation Graduate School of Engineering, Tohoku University
4th Author's Name Shozo MAKINO
4th Author's Affiliation Graduate School of Engineering, Tohoku University
Date 2008-12-09
Paper # NLC2008-32,SP2008-87
Volume (vol) vol.108
Number (no) 337
Page pp.pp.-
#Pages 5
Date of Issue