[2507.02965] Concept-based Adversarial Attack: a Probabilistic Perspective

Computer Science > Computer Vision and Pattern Recognition arXiv:2507.02965 (cs) [Submitted on 30 Jun 2025 (v1), last revised 26 Feb 2026 (this version, v2)] Title:Concept-based Adversarial Attack: a Probabilistic Perspective Authors:Andi Zhang, Xuan Ding, Steven McDonagh, Samuel Kaski View a PDF of the paper titled Concept-based Adversarial Attack: a Probabilistic Perspective, by Andi Zhang and 3 other authors View PDF HTML (experimental) Abstract:We propose a concept-based adversarial attack framework that extends beyond single-image perturbations by adopting a probabilistic perspective. Rather than modifying a single image, our method operates on an entire concept - represented by a distribution - to generate diverse adversarial examples. Preserving the concept is essential, as it ensures that the resulting adversarial images remain identifiable as instances of the original underlying category or identity. By sampling from this concept-based adversarial distribution, we generate images that maintain the original concept but vary in pose, viewpoint, or background, thereby misleading the classifier. Mathematically, this framework remains consistent with traditional adversarial attacks in a principled manner. Our theoretical and empirical results demonstrate that concept-based adversarial attacks yield more diverse adversarial examples and effectively preserve the underlying concept, while achieving higher attack efficiency. Subjects: Computer Vision and Pattern Recognitio...