CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning
Proceedings of the AAAI Conference on Artificial Intelligence · 26.06.2023 · 37 (2) · 1530-1538
DOI: 10.1609/aaai.v37i2.25239 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 26.06.2023. Verfasst von Zihan Li und 4 weiteren Autor:innen. Die Arbeit wurde bisher 11-mal zitiert (Crossref).
Abstract (Originalsprache)
Despite the excellent performance, deep neural networks (DNNs) have been shown to be vulnerable to adversarial examples. Besides, these examples are often transferable among different models. In other words, the same adversarial example can fool multiple models with different architectures at the same time. Based on this property, many black-box transfer-based attack techniques have been developed. However, current transfer-based attacks generally focus on the cross-architecture setting, where the attacker has access to the training data of the target model, which is not guaranteed in realistic situations. In this paper, we design a Cross-Domain Transfer-Based Attack (CDTA), which works in the cross-domain scenario. In this setting, attackers have no information about the target model, such as its architecture and training data. Specifically, we propose a contrastive spectral training method to train a feature extractor on a source domain (e.g., ImageNet) and use it to craft adversarial examples on target domains (e.g., Oxford 102 Flower). Our method corrupts the semantic information of the benign image by scrambling the outputs of both the intermediate feature layers and the final layer of the feature extractor. We evaluate CDTA with 16 target deep models on four datasets with widely varying styles. The results confirm that, in terms of the attack success rate, our approach can consistently outperform the state-of-the-art baselines by an average of 11.45% across all target models. Our code is available at https://github.com/LiulietLee/CDTA.
Zitieren
APA
Li, Z., Wu, W., Su, Y., Zheng, Z., & Lyu, M. R. (2023). CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning. Proceedings of the AAAI Conference on Artificial Intelligence, 37(2), 1530-1538. https://doi.org/10.1609/aaai.v37i2.25239
Vancouver
Li Z, Wu W, Su Y, Zheng Z, Lyu MR. CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning. Proceedings of the AAAI Conference on Artificial Intelligence. 2023;37(2):1530-1538. doi:10.1609/aaai.v37i2.25239
BibTeX
@article{Li2023,
author = {Li, Zihan and Wu, Weibin and Su, Yuxin and Zheng, Zibin and Lyu, Michael R.},
title = {CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2023},
volume = {37},
number = {2},
pages = {1530--1538},
doi = {10.1609/aaai.v37i2.25239}
}
EUVIMED