An efficient method against adversarial attacks

Hongyi Wei; Jing Dong

doi:10.1117/12.2672191

21 March 2023 An efficient method against adversarial attacks

Hongyi Wei, Jing Dong

Proceedings Volume 12609, International Conference on Computer Application and Information Security (ICCAIS 2022); 126091O (2023) https://doi.org/10.1117/12.2672191
Event: International Conference on Computer Application and Information Security (ICCAIS 2022), 2022, ONLINE, ONLINE

Abstract

Previous studies to defense against adversarial examples mostly focused on refining the DNN models but have either shown limited success or required expensive computation. In this paper, we introduce a new detection method against adversarial attacks. Since L0 attackers have similar search patterns, to separate clean examples from adversarial examples, we found a new distance measure on output layer. These strategies have low time and computing costs and can be easily complementary to other defenses. Moreover, our method performs well on adversarial noise localization task.

Citation Download Citation

Hongyi Wei and Jing Dong "An efficient method against adversarial attacks", Proc. SPIE 12609, International Conference on Computer Application and Information Security (ICCAIS 2022), 126091O (21 March 2023); https://doi.org/10.1117/12.2672191

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available