Aggregate a posteriori linear regression for speaker adaptation

Chih Hsien Huang*, Jen-Tzung Chien

*Corresponding author for this work

研究成果: Conference contribution

1 引文 斯高帕斯(Scopus)

摘要

In this paper, we present a rapid and discriminative speaker adaptation algorithm for speech recognition. The adaptation paradigm is constructed under the popular linear regression transformation framework. Attractively, we estimate the regression matrices from the speaker-specific adaptation data according to the aggregate a posteriori criterion, which can be expressed in a form of classification error function. The goal of proposed aggregate a posteriori linear regression (AAPLR) turns out to estimate the discriminative linear regression matrices for transformation-based adaptation so that the classification errors can be minimized. Different from minimum classification error linear regression (MCELR), AAPLR algorithm ha closed-form solution to achieve rapid speaker adaptation. The experimental results reveal that AAPLR speaker adaptation does improve speech recognition performance with moderate computational cost compared to the maximum likelihood linear regression (MLLR), maximum a posteriori linear regression (MAPLR) and MCELR.

原文English
主出版物標題2005 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP '05 - Proceedings - Image and Multidimensional Signal Processing Multimedia Signal Processing
發行者Institute of Electrical and Electronics Engineers Inc.
ISBN(列印)0780388747, 9780780388741
DOIs
出版狀態Published - 1 一月 2005
事件2005 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP '05 - Philadelphia, PA, United States
持續時間: 18 三月 200523 三月 2005

出版系列

名字ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
I
ISSN(列印)1520-6149

Conference

Conference2005 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP '05
國家United States
城市Philadelphia, PA
期間18/03/0523/03/05

指紋 深入研究「Aggregate a posteriori linear regression for speaker adaptation」主題。共同形成了獨特的指紋。

引用此