DocumentCode
352333
Title
Robust estimation for rapid speaker adaptation using discounted likelihood techniques
Author
Gunawardana, Asela ; Byrne, William
Author_Institution
Center for Language & Speech Procesing, Johns Hopkins Univ., Baltimore, MD, USA
Volume
2
fYear
2000
fDate
2000
Abstract
The discounted likelihood procedure, which is a robust extension of the usual EM procedure, is presented, and two approximations which lead to two different variants of the usual maximum likelihood linear regression adaptation scheme are introduced. These schemes are shown to robustly estimate speaker adaptation transforms with very little data. The evaluation is carried out on the Switchboard corpus
Keywords
interpolation; maximum likelihood estimation; speech recognition; transforms; EM procedure; MLLR adaptation scheme; Switchboard corpus; approximations; discounted likelihood techniques; maximum likelihood linear regression; rapid speaker adaptation; robust estimation; speaker adaptation transforms; Acoustic testing; Adaptation model; Approximation algorithms; Linear regression; Loudspeakers; Maximum likelihood linear regression; Natural languages; Robustness; Speech processing; System testing;
fLanguage
English
Publisher
ieee
Conference_Titel
Acoustics, Speech, and Signal Processing, 2000. ICASSP '00. Proceedings. 2000 IEEE International Conference on
Conference_Location
Istanbul
ISSN
1520-6149
Print_ISBN
0-7803-6293-4
Type
conf
DOI
10.1109/ICASSP.2000.859127
Filename
859127
Link To Document