EUVIMEDEUVIMED entdecken
Crossref

Breaking the Stealth-Potency Trade-off in Clean-Image Backdoors with Generative Trigger Optimization

Binyan Xu, Fan Yang, Di Tang, Xilin Dai, Kehuan Zhang

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (32) · 27197-27205

DOI: 10.1609/aaai.v40i32.39935 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Binyan Xu und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Clean-image backdoor attacks, which use only label manipulation in training datasets to compromise deep neural networks, pose a significant threat to security-critical applications. A critical flaw in existing methods is that the poison rate required for a successful attack induces a proportional, and thus noticeable, drop in Clean Accuracy (CA), undermining their stealthiness. This paper presents a new paradigm for clean-image attacks that minimizes this accuracy degradation by optimizing the trigger itself. We introduce Generative Clean-Image Backdoors (GCB), a framework that uses a conditional InfoGAN to identify naturally occurring image features that can serve as potent and stealthy triggers. By ensuring these triggers are easily separable from benign task-related features, GCB enables a victim model to learn the backdoor from an extremely small set of poisoned examples, resulting in a CA drop of less than 1%. Our experiments demonstrate GCB's remarkable versatility, successfully adapting to six datasets, five architectures, and four tasks, including the first demonstration of clean-image backdoors in regression and segmentation. GCB also exhibits resilience against most of the existing backdoor defenses.

Zitieren

APA

Xu, B., Yang, F., Tang, D., Dai, X., & Zhang, K. (2026). Breaking the Stealth-Potency Trade-off in Clean-Image Backdoors with Generative Trigger Optimization. Proceedings of the AAAI Conference on Artificial Intelligence, 40(32), 27197-27205. https://doi.org/10.1609/aaai.v40i32.39935

Vancouver

Xu B, Yang F, Tang D, Dai X, Zhang K. Breaking the Stealth-Potency Trade-off in Clean-Image Backdoors with Generative Trigger Optimization. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(32):27197-27205. doi:10.1609/aaai.v40i32.39935

BibTeX
@article{Xu2026,
  author = {Xu, Binyan and Yang, Fan and Tang, Di and Dai, Xilin and Zhang, Kehuan},
  title = {Breaking the Stealth-Potency Trade-off in Clean-Image Backdoors with Generative Trigger Optimization},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {32},
  pages = {27197--27205},
  doi = {10.1609/aaai.v40i32.39935}
}
Kostenlos starten →