Constructing shared-state hidden Markov Models based on a Bayesian approach

Shinji Watanabe, Yasuhiro Minami, Atsushi Nakamura, Naonori Ueda

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Citations (Scopus)

Abstract

In this paper, we propose a method for constructing sharedstate triphone HMMs (SST-HMMs) within a practical Bayesian framework. In our method, Bayesian model selection criterion is derived for SST-HMM based on the Variational Bayesian approach. The appropriate phonetic decision tree structure of SST-HMM is found by using the criterion according to a given data set. This criterion, unlike the conventional MDL criterion, is applicable even in the case of insufficient amounts of data. We conduct two experiments on speaker independent word recognition in order to prove the effectiveness of the proposed method. The first experiment demonstrates that the Bayesian approach is valid for determining the tree structure. The second experiment demonstrates that the Bayesian criterion can design SST-HMMs with higher recognition performance than the MDL criterion when dealing with small amounts of data.

Original languageEnglish
Title of host publication7th International Conference on Spoken Language Processing, ICSLP 2002
PublisherInternational Speech Communication Association
Pages2669-2672
Number of pages4
Publication statusPublished - 2002
Externally publishedYes
Event7th International Conference on Spoken Language Processing, ICSLP 2002 - Denver, United States
Duration: 2002 Sep 162002 Sep 20

Other

Other7th International Conference on Spoken Language Processing, ICSLP 2002
CountryUnited States
CityDenver
Period02/9/1602/9/20

Fingerprint

experiment
phonetics
Hidden Markov Model
performance
Experiment
Bayesian Model
Decision Tree
Model Selection
Word Recognition
Conventional

ASJC Scopus subject areas

  • Language and Linguistics
  • Linguistics and Language

Cite this

Watanabe, S., Minami, Y., Nakamura, A., & Ueda, N. (2002). Constructing shared-state hidden Markov Models based on a Bayesian approach. In 7th International Conference on Spoken Language Processing, ICSLP 2002 (pp. 2669-2672). International Speech Communication Association.

Constructing shared-state hidden Markov Models based on a Bayesian approach. / Watanabe, Shinji; Minami, Yasuhiro; Nakamura, Atsushi; Ueda, Naonori.

7th International Conference on Spoken Language Processing, ICSLP 2002. International Speech Communication Association, 2002. p. 2669-2672.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Watanabe, S, Minami, Y, Nakamura, A & Ueda, N 2002, Constructing shared-state hidden Markov Models based on a Bayesian approach. in 7th International Conference on Spoken Language Processing, ICSLP 2002. International Speech Communication Association, pp. 2669-2672, 7th International Conference on Spoken Language Processing, ICSLP 2002, Denver, United States, 02/9/16.
Watanabe S, Minami Y, Nakamura A, Ueda N. Constructing shared-state hidden Markov Models based on a Bayesian approach. In 7th International Conference on Spoken Language Processing, ICSLP 2002. International Speech Communication Association. 2002. p. 2669-2672
Watanabe, Shinji ; Minami, Yasuhiro ; Nakamura, Atsushi ; Ueda, Naonori. / Constructing shared-state hidden Markov Models based on a Bayesian approach. 7th International Conference on Spoken Language Processing, ICSLP 2002. International Speech Communication Association, 2002. pp. 2669-2672
@inproceedings{8b6cb3932cb847f6831b0741cc143e2b,
title = "Constructing shared-state hidden Markov Models based on a Bayesian approach",
abstract = "In this paper, we propose a method for constructing sharedstate triphone HMMs (SST-HMMs) within a practical Bayesian framework. In our method, Bayesian model selection criterion is derived for SST-HMM based on the Variational Bayesian approach. The appropriate phonetic decision tree structure of SST-HMM is found by using the criterion according to a given data set. This criterion, unlike the conventional MDL criterion, is applicable even in the case of insufficient amounts of data. We conduct two experiments on speaker independent word recognition in order to prove the effectiveness of the proposed method. The first experiment demonstrates that the Bayesian approach is valid for determining the tree structure. The second experiment demonstrates that the Bayesian criterion can design SST-HMMs with higher recognition performance than the MDL criterion when dealing with small amounts of data.",
author = "Shinji Watanabe and Yasuhiro Minami and Atsushi Nakamura and Naonori Ueda",
year = "2002",
language = "English",
pages = "2669--2672",
booktitle = "7th International Conference on Spoken Language Processing, ICSLP 2002",
publisher = "International Speech Communication Association",

}

TY - GEN

T1 - Constructing shared-state hidden Markov Models based on a Bayesian approach

AU - Watanabe, Shinji

AU - Minami, Yasuhiro

AU - Nakamura, Atsushi

AU - Ueda, Naonori

PY - 2002

Y1 - 2002

N2 - In this paper, we propose a method for constructing sharedstate triphone HMMs (SST-HMMs) within a practical Bayesian framework. In our method, Bayesian model selection criterion is derived for SST-HMM based on the Variational Bayesian approach. The appropriate phonetic decision tree structure of SST-HMM is found by using the criterion according to a given data set. This criterion, unlike the conventional MDL criterion, is applicable even in the case of insufficient amounts of data. We conduct two experiments on speaker independent word recognition in order to prove the effectiveness of the proposed method. The first experiment demonstrates that the Bayesian approach is valid for determining the tree structure. The second experiment demonstrates that the Bayesian criterion can design SST-HMMs with higher recognition performance than the MDL criterion when dealing with small amounts of data.

AB - In this paper, we propose a method for constructing sharedstate triphone HMMs (SST-HMMs) within a practical Bayesian framework. In our method, Bayesian model selection criterion is derived for SST-HMM based on the Variational Bayesian approach. The appropriate phonetic decision tree structure of SST-HMM is found by using the criterion according to a given data set. This criterion, unlike the conventional MDL criterion, is applicable even in the case of insufficient amounts of data. We conduct two experiments on speaker independent word recognition in order to prove the effectiveness of the proposed method. The first experiment demonstrates that the Bayesian approach is valid for determining the tree structure. The second experiment demonstrates that the Bayesian criterion can design SST-HMMs with higher recognition performance than the MDL criterion when dealing with small amounts of data.

UR - http://www.scopus.com/inward/record.url?scp=0141852571&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0141852571&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0141852571

SP - 2669

EP - 2672

BT - 7th International Conference on Spoken Language Processing, ICSLP 2002

PB - International Speech Communication Association

ER -