Presentation 2001/10/1
Efficient Reduction of Gaussian Components Using MDL Criterion for Speech Recognition
Koichi Shinoda, Dieu Tran, Ken-ichi Iso,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A method is proposed to reduce the number of Gaussian components in continuous density hidden Markov models(HMMs). As its initial model, the method employs a well-trained, large-sized HMM in which the components of each state's Gaussian mixture probability density function are clustered into a binary tree. For each state, a subset of Gaussian components is chosen from the Gaussian tree on the basis of the minimum description length(MDL) criterion. By varying the penalty coefficient for large size models in the MDL criterion, it is possible to obtain the total number of Gaussian components desired for smaller models. In our experimental evaluations, the proposed method successfully reduced the number of Gaussian components by 75%, with only 1% degradation in recognition accuracy.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) speech recognition / HMM / MDL criterion / tree structure / Gaussian distribution
Paper # SP2001-83,WIT2001-37
Date of Issue

Conference Information
Committee WIT
Conference Date 2001/10/1(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 Well-being Information Technology(WIT)
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Efficient Reduction of Gaussian Components Using MDL Criterion for Speech Recognition
Sub Title (in English)
Keyword(1) speech recognition
Keyword(2) HMM
Keyword(3) MDL criterion
Keyword(4) tree structure
Keyword(5) Gaussian distribution
1st Author's Name Koichi Shinoda
1st Author's Affiliation Multimedia Research Labs, NEC Corporation()
2nd Author's Name Dieu Tran
2nd Author's Affiliation Multimedia Research Labs, NEC Corporation:(Present address)Currently with Cisco Systems
3rd Author's Name Ken-ichi Iso
3rd Author's Affiliation Multimedia Research Labs, NEC Corporation
Date 2001/10/1
Paper # SP2001-83,WIT2001-37
Volume (vol) vol.101
Number (no) 353
Page pp.pp.-
#Pages 8
Date of Issue