On-line Bayes adaptation of SCHMM parameters for speech recognition

Author

Huo, Qiang ; Chan, Chorkin

Author_Institution

Dept. of Comput. Sci., Hong Kong Univ., Hong Kong

Volume

1

fYear

1995

fDate

9-12 May 1995

Firstpage

708

Abstract

On-line adaptation of semi-continuous (or tied mixture) hidden Markov model (SCHMM) is studied. A theoretical formulation of the segmental quasi-Bayes learning of the mixture coefficients in SCHMM for speech recognition is presented. The practical issues related to the use of this algorithm for on-line speaker adaptation are addressed. A pragmatic on-line adaptation approach to combine the long-term adaptation of the mixture coefficients and the short-term adaptation of the mean vectors of the Gaussian mixture components are also proposed. The viability of these techniques are confirmed in a series of comparative experiments using a 26-word English alphabet vocabulary

Keywords

Bayes methods; Gaussian processes; adaptive signal processing; hidden Markov models; online operation; parameter estimation; speech processing; speech recognition; English alphabet vocabulary; Gaussian mixture components; SCHMM parameters; comparative experiments; long-term adaptation; mean vectors; mixture coefficients; on-line Bayes adaptation; on-line speaker adaptation; segmental quasi-Bayes learning; semi-continuous hidden Markov model; short-term adaptation; speech recognition; theoretical formulation; tied mixture hidden Markov model; Acoustic applications; Acoustic testing; Bayesian methods; Bridges; Computer science; Hidden Markov models; Speech recognition; System testing; Training data; Vocabulary;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1995. ICASSP-95., 1995 International Conference on

Conference_Location

Detroit, MI

ISSN

1520-6149

Print_ISBN

0-7803-2431-5

Type

conf

DOI

10.1109/ICASSP.1995.479792

Filename

479792