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 2021-03-03 14:40
Semantic and Quantitative Explanation for Networks using Feature Interaction
Bohui Xia, Xueting Wang, Toshihiko Yamasaki (The Univ. of Tokyo) IMQ2020-35 IE2020-75 MVE2020-67
Abstract (in Japanese) (See Japanese page) 
(in English) Deep learning-based methods have shown remarkable performances in many tasks. However, it is hard for us to interpret the models, hence explainability of networks have been explored actively. In this work, we propose a method to generate semantic and quantitative explanations using attribute interactions based on an existing model. More specifically, we consider not only contributions from each attribute but also their interactions to prevent low explanatory performance. We test our model on multiple datasets and demonstrate how our method works in real world prediction problems.
Keyword (in Japanese) (See Japanese page) 
(in English) Deep neural networks / Explainability / Semantic attributes / Feature interaction / / / /  
Reference Info. IEICE Tech. Rep., vol. 120, no. 391, MVE2020-67, pp. 127-132, March 2021.
Paper # MVE2020-67 
Date of Issue 2021-02-22 (IMQ, IE, MVE) 
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 IMQ2020-35 IE2020-75 MVE2020-67

Conference Information
Committee MVE IMQ IE CQ  
Conference Date 2021-03-01 - 2021-03-03 
Place (in Japanese) (See Japanese page) 
Place (in English) Online 
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To MVE 
Conference Code 2021-03-MVE-IMQ-IE-CQ 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Semantic and Quantitative Explanation for Networks using Feature Interaction 
Sub Title (in English)  
Keyword(1) Deep neural networks  
Keyword(2) Explainability  
Keyword(3) Semantic attributes  
Keyword(4) Feature interaction  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Bohui Xia  
1st Author's Affiliation The University of Tokyo (The Univ. of Tokyo)
2nd Author's Name Xueting Wang  
2nd Author's Affiliation The University of Tokyo (The Univ. of Tokyo)
3rd Author's Name Toshihiko Yamasaki  
3rd Author's Affiliation The University of Tokyo (The Univ. of Tokyo)
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 2021-03-03 14:40:00 
Presentation Time 25 minutes 
Registration for MVE 
Paper # IMQ2020-35, IE2020-75, MVE2020-67 
Volume (vol) vol.120 
Number (no) no.389(IMQ), no.390(IE), no.391(MVE) 
Page pp.127-132 
#Pages
Date of Issue 2021-02-22 (IMQ, IE, MVE) 


[Return to Top Page]

[Return to IEICE Web Page]


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