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 2023-06-29 15:10
Selective Inference for DNN-driven Saliency Map
Daiki Miwa (NITech), Vo Nguyen Le Duy (RIKEN), Tomohiro Shiraishi (Nagoya Univ.), Ichiro Takeuchi (Nagoya Univ./RIKEN) NC2023-5 IBISML2023-5
Abstract (in Japanese) (See Japanese page) 
(in English) The usefulness of image classification using DNN models has been confirmed in various fields, but the prediction mechanism of these models cannot be fully understood. Therefore, a type of methods called Saliency Map has been proposed to visualize the explanation of predictions, but issues regarding the reliability of the results have been pointed out. In this study, we consider evaluating the reliability of Saliency Map using the framework of statistical hypothesis testing. Unfortunately, conventional statistical hypothesis testing cannot perform valid tests due to the selection bias caused by DNNs. Therefore, we introduce the concept of selective inference, which has been actively studied in recent years, and propose a valid testing method that can properly control Type I Error Rate.
In this paper, we specifically report the results when the proposed method is applied to saliency maps obtained by Grad-cam, one of the most frequently used Saliency Map techniques.
Keyword (in Japanese) (See Japanese page) 
(in English) Selective Inference / Statistical Hypothesis Testing / Saliency Map / Deep Learning / / / /  
Reference Info. IEICE Tech. Rep., vol. 123, no. 91, IBISML2023-5, pp. 30-34, June 2023.
Paper # IBISML2023-5 
Date of Issue 2023-06-22 (NC, IBISML) 
ISSN Online edition: ISSN 2432-6380
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 NC2023-5 IBISML2023-5

Conference Information
Conference Date 2023-06-29 - 2023-07-01 
Place (in Japanese) (See Japanese page) 
Place (in English) OIST Conference Center 
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To IBISML 
Conference Code 2023-06-NC-IBISML-BIO-MPS 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Selective Inference for DNN-driven Saliency Map 
Sub Title (in English)  
Keyword(1) Selective Inference  
Keyword(2) Statistical Hypothesis Testing  
Keyword(3) Saliency Map  
Keyword(4) Deep Learning  
1st Author's Name Daiki Miwa  
1st Author's Affiliation Nagoya Institute of Technology (NITech)
2nd Author's Name Vo Nguyen Le Duy  
2nd Author's Affiliation RIKEN (RIKEN)
3rd Author's Name Tomohiro Shiraishi  
3rd Author's Affiliation Nagoya University (Nagoya Univ.)
4th Author's Name Ichiro Takeuchi  
4th Author's Affiliation Nagoya University/RIKEN (Nagoya Univ./RIKEN)
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 2023-06-29 15:10:00 
Presentation Time 25 minutes 
Registration for IBISML 
Paper # NC2023-5, IBISML2023-5 
Volume (vol) vol.123 
Number (no) no.90(NC), no.91(IBISML) 
Page pp.30-34 
Date of Issue 2023-06-22 (NC, IBISML) 

[Return to Top Page]

[Return to IEICE Web Page]

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