Simultaneous speaker normalisation and utterance labelling using Bayesian/neural net techniques

Author

Cox, S.J. ; Bridle, J.S.

Author_Institution

British Telecom Res. Lab., Ipswich, UK

fYear

1990

fDate

3-6 Apr 1990

Firstpage

161

Abstract

A particular form of neural network is described which has terminals for acoustic patterns, class labels, and speaker parameters. A method of training this network to tune in the speaker parameters to a new speaker is outlined. This process can also be viewed from a Bayesian perspective as maximizing the likelihood of the speaker´s data by optimizing the model and speaker parameters. A method for doing this when the data are labeled is described. Results of using this technique with whole-word hidden Markov models (HMMs) indicate an improvement over speaker-independent performance and, for unlabeled data, a performance close to that achieved on labeled data

Keywords

Bayes methods; Markov processes; adaptive systems; learning systems; neural nets; speech recognition; Bayesian techniques; acoustic patterns; class labels; likelihood-weighted adaptation; network training; neural network; speaker normalisation; speaker parameters; speech recognition; unlabeled data; utterance labelling; whole-word hidden Markov models; Automatic speech recognition; Bayesian methods; Hidden Markov models; Labeling; Laboratories; Loudspeakers; Neural networks; Optimization methods; Radar; Telecommunications;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1990. ICASSP-90., 1990 International Conference on

Conference_Location

Albuquerque, NM

ISSN

1520-6149

Type

conf

DOI

10.1109/ICASSP.1990.115563

Filename

115563