EUVIMEDEUVIMED entdecken
Crossref

AUCPro: AUC-Oriented Provable Robustness Learning

Shilong Bao, Qianqian Xu, Zhiyong Yang, Yuan He, Xiaochun Cao, Qingming Huang

IEEE Transactions on Pattern Analysis and Machine Intelligence · 2025-06 · 47 (6) · 4579-4596

DOI: 10.1109/tpami.2025.3545639 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025-06. Verfasst von Shilong Bao und 5 weiteren Autor:innen. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 114 Literaturangaben. Gefördert von: National Key R&D Program of China, National Natural Science Foundation of China, Youth Innovation Promotion Association of the Chinese Academy of Sciences.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The current studies of provable robustness for deep neural networks (DNNs) usually assume that the class distribution is overall balanced. However, in real-world applications especially for safety-sensitive systems, the class distribution often exhibits a long-tailed property. It is well-known that the Area Under the ROC Curve (AUC) is a more proper metric for long-tailed learning problems. Motivated by this fact, an AUC-oriented provable robustness learning framework (named AUCPro) is first proposed in this paper. The key is to construct a proxy model smoothed by the isotropic Gaussian noise and then consider optimizing the proxy model from the AUC-oriented learning point of view. Theoretically, we provide a certified safety region for AUCPro within which the model would be free from the $\ell _{2}$ℓ2 adversarial attacks. Most importantly, we propose a novel standard to theoretically study the robustness generalization toward unseen data for provable robustness learning approaches. To the best of our knowledge, such a problem remains barely considered in the machine learning community. To be specific, under a general principle for performance-robustness trade-off, we prove that the generalization ability of the resulting model could be equivalently expressed as the expected adversarial risk of AUC under $\ell _{2}$ℓ2 perturbation. On top of this, we present two practical settings to explore the excess risk formed by the difference between the empirical risk of AUCPro and the derived generalization performance. Finally, comprehensive experiments speak to the efficacy of our proposed algorithm.

Förderung

  • National Key R&D Program of China
  • National Natural Science Foundation of China
  • Youth Innovation Promotion Association of the Chinese Academy of Sciences
  • Strategic Priority Research Program of the Chinese Academy of Sciences
  • Postdoctoral Fellowship Program of CPSF

Zitieren

APA

Bao, S., Xu, Q., Yang, Z., He, Y., Cao, X., & Huang, Q. (2025). AUCPro: AUC-Oriented Provable Robustness Learning. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(6), 4579-4596. https://doi.org/10.1109/tpami.2025.3545639

Vancouver

Bao S, Xu Q, Yang Z, He Y, Cao X, Huang Q. AUCPro: AUC-Oriented Provable Robustness Learning. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2025;47(6):4579-4596. doi:10.1109/tpami.2025.3545639

BibTeX
@article{Bao2025,
  author = {Bao, Shilong and Xu, Qianqian and Yang, Zhiyong and He, Yuan and Cao, Xiaochun and Huang, Qingming},
  title = {AUCPro: AUC-Oriented Provable Robustness Learning},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2025},
  volume = {47},
  number = {6},
  pages = {4579--4596},
  doi = {10.1109/tpami.2025.3545639}
}
Kostenlos starten →