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 2018-07-26 17:15
Knowledge Distillation from Neural Network Based Acoustic Model based on Different Decision Tree
Takashi Fukuda, Samuel Thomas (IBM) SP2018-20
Abstract (in Japanese) (See Japanese page) 
(in English) This paper proposes a method to transfer acoustic knowledge from teacher network with a different decision tree to a student network. The teacher model has different output layer from the student network but has high recognition performance. In the proposed method, (1) phone alignments are generated from each phoneme context dependent decision tree, which relates to the output layer of the network, and (2) create a confusion matrix representing a relationship of phoneme contexts between teacher's and student's output nodes. In the experiments, we show that the proposed method contributes to the improvement of student network and report that 9.6% relative improvement was obtained by the proposed method over the acoustic model constructed only with hard labels on noisy environment speech recognition task with Aurora 4.
Keyword (in Japanese) (See Japanese page) 
(in English) Speech recognition / acoustic model / knowledge distillation / decision tree / phone mapping / / /  
Reference Info. IEICE Tech. Rep., vol. 118, no. 160, SP2018-20, pp. 21-24, July 2018.
Paper # SP2018-20 
Date of Issue 2018-07-19 (SP) 
ISSN 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 SP2018-20

Conference Information
Committee SP IPSJ-SLP  
Conference Date 2018-07-26 - 2018-07-27 
Place (in Japanese) (See Japanese page) 
Place (in English) Sago-Royal-Hotel (Hamamatsu) 
Topics (in Japanese) (See Japanese page) 
Topics (in English) Speech recognition and understanding, dialog system, etc. 
Paper Information
Registration To SP 
Conference Code 2018-07-SP-SLP 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Knowledge Distillation from Neural Network Based Acoustic Model based on Different Decision Tree 
Sub Title (in English)  
Keyword(1) Speech recognition  
Keyword(2) acoustic model  
Keyword(3) knowledge distillation  
Keyword(4) decision tree  
Keyword(5) phone mapping  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Takashi Fukuda  
1st Author's Affiliation IBM Japan (IBM)
2nd Author's Name Samuel Thomas  
2nd Author's Affiliation IBM T. J. Watson Research Center (IBM)
3rd Author's Name  
3rd Author's Affiliation ()
4th Author's Name  
4th Author's Affiliation ()
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 ()
Speaker Author-1 
Date Time 2018-07-26 17:15:00 
Presentation Time 30 minutes 
Registration for SP 
Paper # SP2018-20 
Volume (vol) vol.118 
Number (no) no.160 
Page pp.21-24 
#Pages
Date of Issue 2018-07-19 (SP) 


[Return to Top Page]

[Return to IEICE Web Page]


The Institute of Electronics, Information and Communication Engineers (IEICE), Japan