Paper Abstract and Keywords |
Presentation |
2021-10-09 09:30
Explaining Adversarial Examples by the Embedding Structure of Data Manifold Hajime Tasaki, Yuji Kaneko, Jinhui Chao (Chuo Univ.) PRMU2021-19 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
It is widely known that adversarial examples cause misclassification in classifiers using deep learning. Inspite of numerous previous research, the phenomenon has not been well understood and clearly explained, hence there is no effective countermeasures available now. In this research, instead of exploring on neural networks and their training algorithms, we focus on training data of the classifier, as a submanifold embedded in Euclidean space. We show that the occurrence of adversarial examples is due to the embedded structure of the data manifold in the data space. Particularly they lie in the complementary spaces of the tangent spaces of the data manifold. The theory is then verified by generation of adversarial examples and analysis of their distribution. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Adversarial Example / Deep Learning / Data Manifold / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 121, no. 192, PRMU2021-19, pp. 17-21, Oct. 2021. |
Paper # |
PRMU2021-19 |
Date of Issue |
2021-10-01 (PRMU) |
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 |
PRMU2021-19 |
Conference Information |
Committee |
PRMU |
Conference Date |
2021-10-08 - 2021-10-09 |
Place (in Japanese) |
(See Japanese page) |
Place (in English) |
Online |
Topics (in Japanese) |
(See Japanese page) |
Topics (in English) |
Processes and technologies to make research more efficient |
Paper Information |
Registration To |
PRMU |
Conference Code |
2021-10-PRMU |
Language |
Japanese |
Title (in Japanese) |
(See Japanese page) |
Sub Title (in Japanese) |
(See Japanese page) |
Title (in English) |
Explaining Adversarial Examples by the Embedding Structure of Data Manifold |
Sub Title (in English) |
|
Keyword(1) |
Adversarial Example |
Keyword(2) |
Deep Learning |
Keyword(3) |
Data Manifold |
Keyword(4) |
|
Keyword(5) |
|
Keyword(6) |
|
Keyword(7) |
|
Keyword(8) |
|
1st Author's Name |
Hajime Tasaki |
1st Author's Affiliation |
Chuo University (Chuo Univ.) |
2nd Author's Name |
Yuji Kaneko |
2nd Author's Affiliation |
Chuo University (Chuo Univ.) |
3rd Author's Name |
Jinhui Chao |
3rd Author's Affiliation |
Chuo University (Chuo Univ.) |
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 |
2021-10-09 09:30:00 |
Presentation Time |
15 minutes |
Registration for |
PRMU |
Paper # |
PRMU2021-19 |
Volume (vol) |
vol.121 |
Number (no) |
no.192 |
Page |
pp.17-21 |
#Pages |
5 |
Date of Issue |
2021-10-01 (PRMU) |
|