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 2020-03-17 09:45
Rule Extraction from convolutional neural networks -- Rule Extraction with A Pedagogical Method --
Yuya Sato (Tokyo Denki Univ), Hiroshi Tsukimoto (Tokyo Denki Univ.) PRMU2019-89
Abstract (in Japanese) (See Japanese page) 
(in English) We presented a decompositional method of rule extraction from convolutional neural networks in the past. However, we found that the accuracy was extremely low in the case of fully connected layers with hidden layers. This paper presents a pedagogical extracting method. Also, this paper presents a method merging rules and images maximizing the output of convolutional/pooling layers, and visualizing features of the rules extracted from fully connected layers. These methods were applied to convolutional neural networks trained with MNIST dataset.
Keyword (in Japanese) (See Japanese page) 
(in English) convolutional neural network / feature extraction / rule extraction / SmoothGrad / MNIST / pedagogical method / /  
Reference Info. IEICE Tech. Rep., vol. 119, no. 481, PRMU2019-89, pp. 121-126, March 2020.
Paper # PRMU2019-89 
Date of Issue 2020-03-09 (PRMU) 
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 PRMU2019-89

Conference Information
Committee PRMU IPSJ-CVIM  
Conference Date 2020-03-16 - 2020-03-17 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To PRMU 
Conference Code 2020-03-PRMU-CVIM 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Rule Extraction from convolutional neural networks 
Sub Title (in English) Rule Extraction with A Pedagogical Method 
Keyword(1) convolutional neural network  
Keyword(2) feature extraction  
Keyword(3) rule extraction  
Keyword(4) SmoothGrad  
Keyword(5) MNIST  
Keyword(6) pedagogical method  
Keyword(7)  
Keyword(8)  
1st Author's Name Yuya Sato  
1st Author's Affiliation Tokyo Denki University (Tokyo Denki Univ)
2nd Author's Name Hiroshi Tsukimoto  
2nd Author's Affiliation Tokyo Denki University (Tokyo Denki Univ.)
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
Date Time 2020-03-17 09:45:00 
Presentation Time 15 
Registration for PRMU 
Paper # IEICE-PRMU2019-89 
Volume (vol) IEICE-119 
Number (no) no.481 
Page pp.121-126 
#Pages IEICE-6 
Date of Issue IEICE-PRMU-2020-03-09 


[Return to Top Page]

[Return to IEICE Web Page]


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