MEGA Hub

Learning What Not to Learn: Adversarial Disentangled Prompt Tuning for Robust Vision-Language Models

Authors

Do you know Yang Chen?You can claim authorship or link another user.Do you know Zhan Zhuang?You can claim authorship or link another user.Do you know Yanbin Wei?You can claim authorship or link another user.Do you know Zebin Chen?You can claim authorship or link another user.Do you know Hua Liu?You can claim authorship or link another user.Do you know Yu Zhang?You can claim authorship or link another user.

Abstract

While adversarial prompt tuning can enhance robustness of vision-language models efficiently, we find that existing methods aggravate robust generalization overfitting on seen classes, leading to a rapid degradation in performance against adversarial examples of unseen classes as training progresses. We empirically identify that this degradation stems from the tendency of the model to learn pseudo-robust features (i.e., non-generalizable shortcuts). To mitigate this, we propose ADAPT (Adversarial Disentangled Prompt Tuning), a robust prompt tuning framework following the philosophy of ``Learning What Not to Learn''. Specifically, ADAPT uses a dual-prompt mechanism with a target prompt and a pool of decoy prompts. During training, the decoy prompts are guided to entrap diverse pseudo-robust features, while the target prompt is constrained to be orthogonal to the decoys in the embedding space to learn robust features. By disentangling the robust features from the pseudo-robust features, ADAPT effectively prevents robust generalization overfitting. We further provide an analysis showing that the orthogonal loss bounds the effect of shifts in pseudo-robust features on unseen classes, yielding a testing error guarantee. Empirically, extensive experiments demonstrate that ADAPT substantially improves the robustness of the target prompt on unseen classes. The code is available at https://github.com/cheny02/ADAPT-ACMMM2026.

Community

00