Sökning: L773:2162 237X >
Guidance Through Su...
Guidance Through Surrogate: Toward a Generic Diagnostic Attack
-
- Naseer, Muzammal (författare)
- Mohamed Bin Zayed Univ Artificial Intelligence, U Arab Emirates; Australian Natl Univ, Australia
-
- Khan, Salman (författare)
- Mohamed Bin Zayed Univ Artificial Intelligence, U Arab Emirates; Australian Natl Univ, Australia
-
- Porikli, Fatih (författare)
- Qualcomm, CA 92121 USA
-
visa fler...
-
- Khan, Fahad (författare)
- Linköpings universitet,Datorseende,Tekniska fakulteten,Mohamed Bin Zayed Univ Artificial Intelligence, U Arab Emirates
-
visa färre...
-
(creator_code:org_t)
- 2022
- 2022
- Engelska.
-
Ingår i: IEEE Transactions on Neural Networks and Learning Systems. - : IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC. - 2162-237X .- 2162-2388.
- Relaterad länk:
-
https://urn.kb.se/re...
-
visa fler...
-
https://doi.org/10.1...
-
visa färre...
Abstract
Ämnesord
Stäng
- Adversarial training (AT) is an effective approach to making deep neural networks robust against adversarial attacks. Recently, different AT defenses are proposed that not only maintain a high clean accuracy but also show significant robustness against popular and well-studied adversarial attacks, such as projected gradient descent (PGD). High adversarial robustness can also arise if an attack fails to find adversarial gradient directions, a phenomenon known as "gradient masking." In this work, we analyze the effect of label smoothing on AT as one of the potential causes of gradient masking. We then develop a guided mechanism to avoid local minima during attack optimization, leading to a novel attack dubbed guided projected gradient attack (G-PGA). Our attack approach is based on a "match and deceive" loss that finds optimal adversarial directions through guidance from a surrogate model. Our modified attack does not require random restarts a large number of attack iterations or a search for optimal step size. Furthermore, our proposed G-PGA is generic, thus it can be combined with an ensemble attack strategy as we demonstrate in the case of auto-attack, leading to efficiency and convergence speed improvements. More than an effective attack, G-PGA can be used as a diagnostic tool to reveal elusive robustness due to gradient masking in adversarial defenses.
Ämnesord
- NATURVETENSKAP -- Data- och informationsvetenskap -- Datavetenskap (hsv//swe)
- NATURAL SCIENCES -- Computer and Information Sciences -- Computer Sciences (hsv//eng)
Nyckelord
- Smoothing methods; Robustness; Training; Optimization; Behavioral sciences; Computational modeling; Perturbation methods; Adversarial attack; gradient masking; guided optimization; image classification; label smoothing
Publikations- och innehållstyp
- ref (ämneskategori)
- art (ämneskategori)
Hitta via bibliotek
Till lärosätets databas