HKUST Library Institutional Repository Banner

HKUST Institutional Repository >
Computer Science and Engineering >
CSE Conference Papers >

Please use this identifier to cite or link to this item:
Title: Robustness of several kernel-based fast adaptation methods on noisy LVCSR
Authors: Mak, Brian Kan-Wing
Hsiao, Roger
Keywords: Fast adaptation
Kernel method
Kernel eigenspace-based MLLR
Embedded kernel eigenvoice
Reference speaker weighting
Issue Date: 2007
Citation: Proceedings of the 8th Annual Conference of the International Speech Communication Association ( Interspeech 2007), 27-31 August 2007, Antwerp, Belgium, p. 266-269
Abstract: We have been investigating the use of kernel methods to improve conventional linear adaptation algorithms for fast adaptation, when there are less than 10s of adaptation speech. On clean speech, we had shown that our new kernel-based adaptation methods, namely, embedded kernel eigenvoice (eKEV) and kernel eigenspace-based MLLR (KEMLLR) outperformed their linear counterparts. In this paper, we study their unsupervised adaptation performance under additive and convoluted noises using the Aurora4 Corpus, with no assumption or prior knowledge of the noise type and its level. It is found that both eKEV and KEMLLR adaptation continue to outperform MAP and MLLR, and the simple reference speaker weighting (RSW) algorithm continues to perform favorably with KEMLLR. Furthermore, KEMLLR adaptation gives the greatest overall improvement over the speaker-independent model by about 19%.
Rights: We would like to give credit to International Speech Communication Association for granting us permission to repost this article.
Appears in Collections:CSE Conference Papers

Files in This Item:

File Description SizeFormat
is2007kadapt.pdfpre-published version60KbAdobe PDFView/Open

All items in this Repository are protected by copyright, with all rights reserved.