Generalizable Black-Box Adversarial Attack With Meta Learning
IEEE Transactions on Pattern Analysis and Machine Intelligence · 2024-03 · 46 (3) · 1804-1818
DOI: 10.1109/tpami.2022.3194988 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024-03. Verfasst von Fei Yin und 6 weiteren Autor:innen. Die Arbeit wurde bisher 54-mal zitiert (Crossref) und nennt 69 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Shenzhen Key Laboratory of Marine IntelliSense and Computation, Shenzhen Science and Technology Program.
Abstract (Originalsprache)
In the scenario of black-box adversarial attack, the target model's parameters are unknown, and the attacker aims to find a successful adversarial perturbation based on query feedback under a query budget. Due to the limited feedback information, existing query-based black-box attack methods often require many queries for attacking each benign example. To reduce query cost, we propose to utilize the feedback information across historical attacks, dubbed example-level adversarial transferability. Specifically, by treating the attack on each benign example as one task, we develop a meta-learning framework by training a meta generator to produce perturbations conditioned on benign examples. When attacking a new benign example, the meta generator can be quickly fine-tuned based on the feedback information of the new task as well as a few historical attacks to produce effective perturbations. Moreover, since the meta-train procedure consumes many queries to learn a generalizable generator, we utilize model-level adversarial transferability to train the meta generator on a white-box surrogate model, then transfer it to help the attack against the target model. The proposed framework with the two types of adversarial transferability can be naturally combined with any off-the-shelf query-based attack methods to boost their performance, which is verified by extensive experiments. The source code is available at https://github.com/SCLBD/MCG-Blackbox.
Förderung
- National Natural Science Foundation of China
- Shenzhen Key Laboratory of Marine IntelliSense and Computation
- Shenzhen Science and Technology Program
- University Development fund of the Chinese University of Hong Kong
- CCF-Tencent Open Fund
Zitieren
APA
Yin, F., Zhang, Y., Wu, B., Feng, Y., Zhang, J., Fan, Y., & Yang, Y. (2024). Generalizable Black-Box Adversarial Attack With Meta Learning. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(3), 1804-1818. https://doi.org/10.1109/tpami.2022.3194988
Vancouver
Yin F, Zhang Y, Wu B, Feng Y, Zhang J, Fan Y, et al. Generalizable Black-Box Adversarial Attack With Meta Learning. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2024;46(3):1804-1818. doi:10.1109/tpami.2022.3194988
BibTeX
@article{Yin2024,
author = {Yin, Fei and Zhang, Yong and Wu, Baoyuan and Feng, Yan and Zhang, Jingyi and Fan, Yanbo and Yang, Yujiu},
title = {Generalizable Black-Box Adversarial Attack With Meta Learning},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
year = {2024},
volume = {46},
number = {3},
pages = {1804--1818},
doi = {10.1109/tpami.2022.3194988}
}
EUVIMED