DocumentCode :
3163026
Title :
Lasso environment model combination for robust speech recognition
Author :
Xiao, Xiong ; Li, Jinyu ; Chng, Eng Siong ; Li, Haizhou
Author_Institution :
Temasek Lab., Nanyang Technol. Univ., Singapore, Singapore
fYear :
2012
fDate :
25-30 March 2012
Firstpage :
4305
Lastpage :
4308
Abstract :
In this paper, we propose a novel acoustic model adaptation method for noise robust speech recognition. Model combination is a common way to adapt acoustic models to a target test environment. For example, the mean supervectors of the adapted model are obtained as a linear combination of mean supervectors of many pre-trained environment-dependent acoustic models. Usually, the combination weights are estimated using a maximum likelihood (ML) criterion and the weights are nonzero for all the mean supervectors. We propose to estimate the weights by using Lasso (least absolute shrinkage and selection operator) which imposes an L1 regularization term in the weight estimation problem to shrink some weights to exactly zero. Our study shows that Lasso usually shrinks to zero the weights of those mean supervectors not relevant to the test environment. By removing some nonrelevant supervectors, the obtained mean supervectors are found to be more robust against noise distortions. Experimental results on Aurora-2 task show that the Lasso-based mean combination consistently outperforms ML-based combination.
Keywords :
regression analysis; speech recognition; Lasso based mean combination; Lasso environment model combination; acoustic model adaptation; combination weights; environment dependent acoustic model; maximum likelihood criterion; mean supervectors; noise robust speech recognition; regularization term; weight estimation problem; Acoustics; Adaptation models; Maximum likelihood estimation; Noise; Training; Vectors; L1 regularization; Lasso regression; model adaptation; model combination; noise robust speech recognition;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Acoustics, Speech and Signal Processing (ICASSP), 2012 IEEE International Conference on
Conference_Location :
Kyoto
ISSN :
1520-6149
Print_ISBN :
978-1-4673-0045-2
Electronic_ISBN :
1520-6149
Type :
conf
DOI :
10.1109/ICASSP.2012.6288871
Filename :
6288871
Link To Document :
بازگشت