{"ID":5346753,"CreatedAt":"2026-06-30T04:09:55.830587294Z","UpdatedAt":"2026-07-02T14:12:34.668891255Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2606.30342","arxiv_id":"2606.30342","title":"A Classifier-Agnostic Zero-Shot Adversarial Attack Detection via CLIP","abstract":"Adversarial attacks pose a challenge to the reliability of deep learning models, motivating effective detection methods. Existing techniques often rely on attack-specific assumptions, access to adversarial samples, or knowledge of the underlying classifier (white-box). We propose \\textit{$A^4D$ (\\textbf{A}ttack- and \\textbf{A}rchitecture-\\textbf{A}gnostic \\textbf{A}dversarial \\textbf{D}etector)}, a completely black-box, zero-shot adversarial attack detection framework that utilizes prompt-based similarity scores derived from CLIP. To the best of our knowledge this is the first attempt to utilize CLIP for such a task. The method is based on two key observations: (i) CLIP is sensitive even to small imperceptible non-semantic perturbations; (ii) The shift in CLIP embedding space is not arbitrary and can be used as a robust attack indicator. Experiments across multiple attacks, datasets and classifiers validate that $A^4D$ achieves SOTA detection results in the attack-agnostic and classifier-agnostic setting.","short_abstract":"Adversarial attacks pose a challenge to the reliability of deep learning models, motivating effective detection methods. Existing techniques often rely on attack-specific assumptions, access to adversarial samples, or knowledge of the underlying classifier (white-box). We propose \\textit{$A^4D$ (\\textbf{A}ttack- and \\t...","url_abs":"https://arxiv.org/abs/2606.30342","url_pdf":"https://arxiv.org/pdf/2606.30342v1","authors":"[\"Hodaya Krakover\",\"Meir Yossef Levi\",\"Eyal Gofer\",\"Guy Gilboa\"]","published":"2026-06-29T14:19:20Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[]","has_code":false}