Dividing the distributions of HMM and linear interpolation in speech recognition

Kiyoshi Asai; Satoru Hayamizu; Ken'ichi Handa

doi:10.1109/ICASSP.1992.225980

Dividing the distributions of HMM and linear interpolation in speech recognition

Kiyoshi Asai, Satoru Hayamizu, Ken'ichi Handa

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

In this paper, we present an explicit criterion for deciding whether to divide the HMM-based phone model into more precise(for example, context dependent) models, or not. We also discuss the cases when linear interpolations of these models are used, and present an explicit solution of the interpolation weight λ. These criteria are obtained by evaluating the estimation errors of the distributions of these models. We show that the estimation errors should be evaluated by the projected covariance matrices of the estimation errors of the logarithm of the probabilities.

Original language	English
Title of host publication	ICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	29-32
Number of pages	4
ISBN (Electronic)	0780305329
DOIs	https://doi.org/10.1109/ICASSP.1992.225980
Publication status	Published - 1992
Externally published	Yes
Event	1992 International Conference on Acoustics, Speech, and Signal Processing, ICASSP 1992 - San Francisco, United States Duration: 1992 Mar 23 → 1992 Mar 26

Publication series

Name	ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume	1
ISSN (Print)	1520-6149

Other

Other	1992 International Conference on Acoustics, Speech, and Signal Processing, ICASSP 1992
Country/Territory	United States
City	San Francisco
Period	92/3/23 → 92/3/26

ASJC Scopus subject areas

Software
Signal Processing
Electrical and Electronic Engineering

Access to Document

10.1109/ICASSP.1992.225980

Cite this

Asai, K., Hayamizu, S., & Handa, K. (1992). Dividing the distributions of HMM and linear interpolation in speech recognition. In ICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing (pp. 29-32). [225980] (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings; Vol. 1). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.1992.225980

Dividing the distributions of HMM and linear interpolation in speech recognition. / Asai, Kiyoshi; Hayamizu, Satoru; Handa, Ken'ichi.
ICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing. Institute of Electrical and Electronics Engineers Inc., 1992. p. 29-32 225980 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings; Vol. 1).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Asai, K, Hayamizu, S & Handa, K 1992, Dividing the distributions of HMM and linear interpolation in speech recognition. in ICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing., 225980, ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, vol. 1, Institute of Electrical and Electronics Engineers Inc., pp. 29-32, 1992 International Conference on Acoustics, Speech, and Signal Processing, ICASSP 1992, San Francisco, United States, 92/3/23. https://doi.org/10.1109/ICASSP.1992.225980

Asai K, Hayamizu S, Handa K. Dividing the distributions of HMM and linear interpolation in speech recognition. In ICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing. Institute of Electrical and Electronics Engineers Inc. 1992. p. 29-32. 225980. (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). doi: 10.1109/ICASSP.1992.225980

Asai, Kiyoshi ; Hayamizu, Satoru ; Handa, Ken'ichi. / Dividing the distributions of HMM and linear interpolation in speech recognition. ICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing. Institute of Electrical and Electronics Engineers Inc., 1992. pp. 29-32 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).

@inproceedings{2638e8db70f147b7beb9b482a77bb144,

title = "Dividing the distributions of HMM and linear interpolation in speech recognition",

abstract = "In this paper, we present an explicit criterion for deciding whether to divide the HMM-based phone model into more precise(for example, context dependent) models, or not. We also discuss the cases when linear interpolations of these models are used, and present an explicit solution of the interpolation weight λ. These criteria are obtained by evaluating the estimation errors of the distributions of these models. We show that the estimation errors should be evaluated by the projected covariance matrices of the estimation errors of the logarithm of the probabilities.",

author = "Kiyoshi Asai and Satoru Hayamizu and Ken'ichi Handa",

note = "Publisher Copyright: {\textcopyright} 1992 IEEE.; 1992 International Conference on Acoustics, Speech, and Signal Processing, ICASSP 1992 ; Conference date: 23-03-1992 Through 26-03-1992",

year = "1992",

doi = "10.1109/ICASSP.1992.225980",

language = "English",

series = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "29--32",

booktitle = "ICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing",

}

TY - GEN

T1 - Dividing the distributions of HMM and linear interpolation in speech recognition

AU - Asai, Kiyoshi

AU - Hayamizu, Satoru

AU - Handa, Ken'ichi

PY - 1992

Y1 - 1992

N2 - In this paper, we present an explicit criterion for deciding whether to divide the HMM-based phone model into more precise(for example, context dependent) models, or not. We also discuss the cases when linear interpolations of these models are used, and present an explicit solution of the interpolation weight λ. These criteria are obtained by evaluating the estimation errors of the distributions of these models. We show that the estimation errors should be evaluated by the projected covariance matrices of the estimation errors of the logarithm of the probabilities.

AB - In this paper, we present an explicit criterion for deciding whether to divide the HMM-based phone model into more precise(for example, context dependent) models, or not. We also discuss the cases when linear interpolations of these models are used, and present an explicit solution of the interpolation weight λ. These criteria are obtained by evaluating the estimation errors of the distributions of these models. We show that the estimation errors should be evaluated by the projected covariance matrices of the estimation errors of the logarithm of the probabilities.

UR - http://www.scopus.com/inward/record.url?scp=84968430260&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84968430260&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.1992.225980

DO - 10.1109/ICASSP.1992.225980

M3 - Conference contribution

AN - SCOPUS:84968430260

T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

SP - 29

EP - 32

BT - ICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 1992 International Conference on Acoustics, Speech, and Signal Processing, ICASSP 1992

Y2 - 23 March 1992 through 26 March 1992

ER -

Dividing the distributions of HMM and linear interpolation in speech recognition

Abstract

Publication series

Other

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this