Bayesian approaches in speech recognition

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

Abstract

This paper focuses on applications of Bayesian approaches to speech recognition. Bayesian approaches have been widely studied in statistics and machine learning fields, and one of the advantages of the Bayesian approaches is to improve generalization ability compared to maximum likelihood approaches. The effectiveness for speech recognition is shown experimentally in speaker adaptation tasks by using Maximum A Posterior (MAP) and model complexity control by using Bayesian Information Criterion (BIC). This paper introduces the variational Bayesian approaches, in addition to the MAP, BIC and other Bayesian techniques, for speech recognition. VBEC (Variational Bayesian Estimation and Clustering for speech recognition) is a fully Bayesian speech recognition framework, and achieves robust acoustic modeling and speech classification. This paper explains the formulation and experimental effectiveness of these Bayesian approaches for speech recognition.

Original languageEnglish
Title of host publicationAPSIPA ASC 2011 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011
Pages170-179
Number of pages10
Publication statusPublished - 2011
Externally publishedYes
EventAsia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011, APSIPA ASC 2011 - Xi'an
Duration: 2011 Oct 182011 Oct 21

Other

OtherAsia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011, APSIPA ASC 2011
CityXi'an
Period11/10/1811/10/21

Fingerprint

Speech recognition
Maximum likelihood
Learning systems
Acoustics
Statistics

ASJC Scopus subject areas

  • Information Systems
  • Signal Processing

Cite this

Watanabe, S. (2011). Bayesian approaches in speech recognition. In APSIPA ASC 2011 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011 (pp. 170-179)

Bayesian approaches in speech recognition. / Watanabe, Shinji.

APSIPA ASC 2011 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011. 2011. p. 170-179.

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

Watanabe, S 2011, Bayesian approaches in speech recognition. in APSIPA ASC 2011 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011. pp. 170-179, Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011, APSIPA ASC 2011, Xi'an, 11/10/18.
Watanabe S. Bayesian approaches in speech recognition. In APSIPA ASC 2011 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011. 2011. p. 170-179
Watanabe, Shinji. / Bayesian approaches in speech recognition. APSIPA ASC 2011 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011. 2011. pp. 170-179
@inproceedings{d964291defd04fd6ac33d9c331ff9f33,
title = "Bayesian approaches in speech recognition",
abstract = "This paper focuses on applications of Bayesian approaches to speech recognition. Bayesian approaches have been widely studied in statistics and machine learning fields, and one of the advantages of the Bayesian approaches is to improve generalization ability compared to maximum likelihood approaches. The effectiveness for speech recognition is shown experimentally in speaker adaptation tasks by using Maximum A Posterior (MAP) and model complexity control by using Bayesian Information Criterion (BIC). This paper introduces the variational Bayesian approaches, in addition to the MAP, BIC and other Bayesian techniques, for speech recognition. VBEC (Variational Bayesian Estimation and Clustering for speech recognition) is a fully Bayesian speech recognition framework, and achieves robust acoustic modeling and speech classification. This paper explains the formulation and experimental effectiveness of these Bayesian approaches for speech recognition.",
author = "Shinji Watanabe",
year = "2011",
language = "English",
pages = "170--179",
booktitle = "APSIPA ASC 2011 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011",

}

TY - GEN

T1 - Bayesian approaches in speech recognition

AU - Watanabe, Shinji

PY - 2011

Y1 - 2011

N2 - This paper focuses on applications of Bayesian approaches to speech recognition. Bayesian approaches have been widely studied in statistics and machine learning fields, and one of the advantages of the Bayesian approaches is to improve generalization ability compared to maximum likelihood approaches. The effectiveness for speech recognition is shown experimentally in speaker adaptation tasks by using Maximum A Posterior (MAP) and model complexity control by using Bayesian Information Criterion (BIC). This paper introduces the variational Bayesian approaches, in addition to the MAP, BIC and other Bayesian techniques, for speech recognition. VBEC (Variational Bayesian Estimation and Clustering for speech recognition) is a fully Bayesian speech recognition framework, and achieves robust acoustic modeling and speech classification. This paper explains the formulation and experimental effectiveness of these Bayesian approaches for speech recognition.

AB - This paper focuses on applications of Bayesian approaches to speech recognition. Bayesian approaches have been widely studied in statistics and machine learning fields, and one of the advantages of the Bayesian approaches is to improve generalization ability compared to maximum likelihood approaches. The effectiveness for speech recognition is shown experimentally in speaker adaptation tasks by using Maximum A Posterior (MAP) and model complexity control by using Bayesian Information Criterion (BIC). This paper introduces the variational Bayesian approaches, in addition to the MAP, BIC and other Bayesian techniques, for speech recognition. VBEC (Variational Bayesian Estimation and Clustering for speech recognition) is a fully Bayesian speech recognition framework, and achieves robust acoustic modeling and speech classification. This paper explains the formulation and experimental effectiveness of these Bayesian approaches for speech recognition.

UR - http://www.scopus.com/inward/record.url?scp=84866873456&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84866873456&partnerID=8YFLogxK

M3 - Conference contribution

SP - 170

EP - 179

BT - APSIPA ASC 2011 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference 2011

ER -