Information: Join today and make your research activities more affordable! Technical workshop participation fees and annual registration fees are available at member rates.
Notice: [Important] Announcement of Changes to Registration Fee Payment and Manuscript Upload Procedures for IEICE Technical Meetings
IEICE Technical Committee Submission System
Conference Paper's Information		 Online Proceedings
[Sign in]
Tech. Rep. Archives
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2010-12-20 17:20
Robust Acoustic Modeling Using MLLR Transformation-based Speech Feature Generation
Arata Itoh, Sunao Hara, Norihide Kitaoka, Kazuya Takeda (Nagoya Univ.) NLC2010-19 SP2010-92
Abstract (in Japanese) (See Japanese page) 
(in English) We propose a novel acoustic model training method based on the new acoustic feature generation. Recently, the speaker adaptation method, such as MLLR and MAP, are widely used. However, all speaker adaptation methods need adaptation data. On the contrary, our method makes speaker-independent acoustic models that cover not only known but also unknown speakers. We focused on MLLR transformation matrix. Our method is a kind of generative training which generates new acoustic features by inverse transformation of MLLR transformation matrix and uses generated features to train acoustic models. We obtain MLLR transformation matrices from a limited number of existing speakers. Then we extract the bases of the MLLR transformation matrices using PCA and express MLLR transformation matrix by linear combination of bases. The probability distribution of the weight parameters to express the MLLR transformation matrices for the existing speakers are estimated. Finally we generate pseudo-speaker MLLR transformation by sampling the weight parameters from the distribution and apply the inverse of the transformation to the normalized existing speaker features to generate the pseudo-speakers' features. Using these features, we train the acoustic models. Evaluation results show that the acoustic models trained by our method are robust for unknown speakers.
Keyword (in Japanese) (See Japanese page) 
(in English) Speech Recognition / Acoustic Model / MLLR Transformation Matrix / Generative Training / / / /  
Reference Info. IEICE Tech. Rep., vol. 110, no. 357, SP2010-92, pp. 55-60, Dec. 2010.
Paper # SP2010-92 
Date of Issue 2010-12-13 (NLC, SP) 
ISSN Print edition: ISSN 0913-5685    Online edition: ISSN 2432-6380
Copyright
and
reproduction		 All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034)
Download PDF NLC2010-19 SP2010-92

Conference Information
Committee NLC SP  
Conference Date 2010-12-20 - 2010-12-21 
Place (in Japanese) (See Japanese page) 
Place (in English) National Olympics Memorial Youth Center 
Topics (in Japanese) (See Japanese page) 
Topics (in English) Information Access, Speech and Language Processing, etc. 
Paper Information
Registration To SP 
Conference Code 2010-12-NLC-SP 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Robust Acoustic Modeling Using MLLR Transformation-based Speech Feature Generation 
Sub Title (in English)  
Keyword(1) Speech Recognition  
Keyword(2) Acoustic Model  
Keyword(3) MLLR Transformation Matrix  
Keyword(4) Generative Training  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Arata Itoh  
1st Author's Affiliation Nagoya University (Nagoya Univ.)
2nd Author's Name Sunao Hara  
2nd Author's Affiliation Nagoya University (Nagoya Univ.)
3rd Author's Name Norihide Kitaoka  
3rd Author's Affiliation Nagoya University (Nagoya Univ.)
4th Author's Name Kazuya Takeda  
4th Author's Affiliation Nagoya University (Nagoya Univ.)
5th Author's Name  
5th Author's Affiliation ()
6th Author's Name  
6th Author's Affiliation ()
7th Author's Name  
7th Author's Affiliation ()
8th Author's Name  
8th Author's Affiliation ()
9th Author's Name  
9th Author's Affiliation ()
10th Author's Name  
10th Author's Affiliation ()
11th Author's Name  
11th Author's Affiliation ()
12th Author's Name  
12th Author's Affiliation ()
13th Author's Name  
13th Author's Affiliation ()
14th Author's Name  
14th Author's Affiliation ()
15th Author's Name  
15th Author's Affiliation ()
16th Author's Name  
16th Author's Affiliation ()
17th Author's Name  
17th Author's Affiliation ()
18th Author's Name  
18th Author's Affiliation ()
19th Author's Name  
19th Author's Affiliation ()
20th Author's Name  
20th Author's Affiliation ()
21st Author's Name  
21st Author's Affiliation ()
22nd Author's Name  
22nd Author's Affiliation ()
23rd Author's Name  
23rd Author's Affiliation ()
24th Author's Name  
24th Author's Affiliation ()
25th Author's Name  
25th Author's Affiliation ()
26th Author's Name / /
26th Author's Affiliation ()
()
27th Author's Name / /
27th Author's Affiliation ()
()
28th Author's Name / /
28th Author's Affiliation ()
()
29th Author's Name / /
29th Author's Affiliation ()
()
30th Author's Name / /
30th Author's Affiliation ()
()
31st Author's Name / /
31st Author's Affiliation ()
()
32nd Author's Name / /
32nd Author's Affiliation ()
()
33rd Author's Name / /
33rd Author's Affiliation ()
()
34th Author's Name / /
34th Author's Affiliation ()
()
35th Author's Name / /
35th Author's Affiliation ()
()
36th Author's Name / /
36th Author's Affiliation ()
()
Speaker Author-1 
Date Time 2010-12-20 17:20:00 
Presentation Time 25 minutes 
Registration for SP 
Paper # NLC2010-19, SP2010-92 
Volume (vol) vol.110 
Number (no) no.356(NLC), no.357(SP) 
Page pp.55-60 
#Pages 6 
Date of Issue 2010-12-13 (NLC, SP) 

[Return to Top Page]

[Return to IEICE Web Page]
The Institute of Electronics, Information and Communication Engineers (IEICE), Japan