English

Closed-Loop Bidirectional Prompting for Adversarial Robustness of Vision Language Models

Computer Vision and Pattern Recognition 2026-05-26 v1

Abstract

Vision Language Models adapt well to downstream tasks but are highly vulnerable to adversarial perturbations that disrupt cross-modal semantic alignment. Existing defenses are largely unidirectional or structural, failing to exploit bidirectional cross-modal complementarity and instance-wise adaptive protection. To overcome the limitations of unidirectional and static defenses in adversarial settings, we propose Closed-Loop Bidirectional Prompting, casting robust adaptation as cross-modal agreement recovery via a dynamic feedback loop on frozen encoders. A Semantic Anchor is introduced as a stable prior to constrain cyclic updates and mitigate perturbation-induced feature corruption. Through anchor-based bootstrapping, textual semantics denoise visual representations, while the refined visuals enable instance-adaptive prompt updating, yielding a rectified and robust consensus. Extensive evaluations across 11 datasets validate state-of-the-art robustness and strong base-to-new generalization, while maintaining a favorable trade-off between computational cost and accuracy.

Keywords

Cite

@article{arxiv.2605.25922,
  title  = {Closed-Loop Bidirectional Prompting for Adversarial Robustness of Vision Language Models},
  author = {Xiao Liu and Jiaxiang Liu and Boci Peng and Boren Hu and Yusong Wang and Xiwen Chen and Prayag Tiwari and Liming Zhang and Mingkun Xu},
  journal= {arXiv preprint arXiv:2605.25922},
  year   = {2026}
}

Comments

24 pages, 8 figures

R2 v1 2026-07-22T07:32:39.805Z