ReACT-CLIP: Response-Aware Test-Time Defense for Vision--Language Models

• URL: http://arxiv.org/abs/2608.01067v2
• Date: Sat, 08 Aug 2026 13:10:26 GMT
• Title: ReACT-CLIP: Response-Aware Test-Time Defense for Vision--Language Models
• Abstract summary: ReACT-CLIP is a test-time defense that determines how strongly each input should be corrected and whether defensive intervention is necessary.<n>It delivers substantial robustness gains across diverse attack types and strengths while largely preserving clean accuracy.
• Score: 34.88445604400548
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Training-free test-time defenses offer a practical way to improve the adversarial robustness of CLIP-style vision--language models without modifying the pretrained model. However, their correction strength is typically fixed for a narrow range of attack budgets, even though the attack budget is unknown at inference and the required correction varies across samples. We show that this mismatch causes existing defenses to degrade sharply as attacks strengthen. We introduce ReACT-CLIP, a response-conditioned test-time defense that separately determines how strongly each input should be corrected and whether defensive intervention is necessary. Our key observation is that the relative increase in CLIP visual-feature drift between low- and high-noise probes provides a graded, sample-specific proxy for correction demand. ReACT-CLIP maps this relative cross-noise drift to the Gaussian noise scale used to construct a stable, noise-averaged feature anchor, enabling the corrective reach to adapt to each input. To determine whether intervention is necessary, we further observe that clean inputs retain stable class-probability distributions under weak spatial augmentations, whereas adversarial inputs exhibit greater variation. ReACT-CLIP quantifies this variation using a prediction-instability score computed by Jensen--Shannon divergence and combines it with relative cross-noise drift to form the defensive intervention score. ReACT-CLIP requires no model or prompt training, and its correction-strength mapping is calibrated once and fixed across datasets and attack budgets. Across 12 downstream datasets, as well as ImageNet and its distribution-shifted variants, ReACT-CLIP delivers substantial robustness gains across diverse attack types and strengths while largely preserving clean accuracy.

Related papers

• CAS: Conformalized Agentic Search via Adaptive Retrieval and Policy Weighting [11.613315225188302]
  We introduce Conformal Prediction (CP) and propose Conformalized Agentic Search (CAS)<n>This framework establishes reliability guarantees on both the retrieval and training sides.<n>Our framework significantly improves reasoning accuracy while drastically reducing redundant tool invocations.
  arXiv  Detail & Related papers  (2026-08-21T06:29:52Z)
• SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute [62.3458279176813]
  Self-Verifying Refinement is an oracle-free multi-turn reinforcement learning framework.<n>It learns to use self-verification as a compute-control policy.<n>It achieves a macro-average accuracy of 0.563 with only 2.99 inference turns on average.
  arXiv  Detail & Related papers  (2026-07-30T16:20:58Z)
• Robustifying Vision-Language Models via Test-Time Prompt Adaptation [23.05168102474528]
  We propose RITA, a test-tIme prompt-TAdaptation framework that shifts from sample-level estimates to distribution-level alignment.<n>Specifically, RITA employs optimal transport to align the distribution of augmented visual features with textual prototypes, mitigating adversarial outliers and rectifying cross-modal semantic misalignment.
  arXiv  Detail & Related papers  (2026-07-10T14:19:42Z)
• AGC: Adaptive Geodesic Correction for Adversarial Robustness on Vision-Language Models [38.50355675955809]
  AGC achieves superior performance across eight fine-grained datasets and three CLIP backbones.<n>Our findings reveal a fundamental geometric property of CLIP features, offering a highly efficient and effective paradigm for robust multimodal deployment.
  arXiv  Detail & Related papers  (2026-05-15T03:48:22Z)
• Multimodal Learning on Low-Quality Data with Conformal Predictive Self-Calibration [72.0672328514289]
  Multimodal learning often grapples with the challenge of low-quality data.<n>We propose a unified framework, termed Conformal Predictive Self-Calibration.<n>Our framework consistently outperforms existing state-of-the-art methods.
  arXiv  Detail & Related papers  (2026-05-05T14:48:52Z)
• Do Not Imitate, Reinforce: Iterative Classification via Belief Refinement [10.794100274863816]
  We propose Reinforced Iterative Classification (RIC), which replaces the imitative objective with Reinforcement Learning (RL)<n>RIC deploys a recurrent agent that iteratively updates a predictive distribution over classes, receiving reward for stepwise improvement in prediction quality.<n>On image classification benchmarks, RIC matches the accuracy of supervised baselines with improved calibration and learns to allocate computation adaptively across inputs.
  arXiv  Detail & Related papers  (2026-04-23T23:06:48Z)
• LATA: Laplacian-Assisted Transductive Adaptation for Conformal Uncertainty in Medical VLMs [61.06744611795341]
  Medical vision-language models (VLMs) are strong zero-shot recognizers for medical imaging.<n>We propose texttttextbfLATA (Laplacian-Assisted Transductive Adaptation), a textittraining- and label-free refinement.<n>texttttextbfLATA sharpens zero-shot predictions without compromising exchangeability.
  arXiv  Detail & Related papers  (2026-02-19T16:45:38Z)
• Enhancing CLIP Robustness via Cross-Modality Alignment [54.01929554563447]
  We propose Cross-modality Alignment, an optimal transport-based framework for vision-language models.<n> COLA restores global image-text alignment and local structural consistency in the feature space.<n> COLA is training-free and compatible with existing fine-tuned models.
  arXiv  Detail & Related papers  (2025-10-28T03:47:44Z)
• Bayesian Test-time Adaptation for Object Recognition and Detection with Vision-language Models [86.53246292425699]
  We present BCA+, a training-free framework for TTA for both object recognition and detection.<n>We formulate adaptation as a Bayesian inference problem, where final predictions are generated by fusing the initial VLM output with a cache-based prediction.<n>BCA+ achieves state-of-the-art performance on both recognition and detection benchmarks.
  arXiv  Detail & Related papers  (2025-10-03T06:27:33Z)
• Learn from the Past: A Proxy Guided Adversarial Defense Framework with Self Distillation Regularization [53.04697800214848]
  Adversarial Training (AT) is pivotal in fortifying the robustness of deep learning models. AT methods, relying on direct iterative updates for target model's defense, frequently encounter obstacles such as unstable training and catastrophic overfitting. We present a general proxy guided defense framework, LAST' (bf Learn from the Pbf ast)
  arXiv  Detail & Related papers  (2023-10-19T13:13:41Z)
• Robust Feature Inference: A Test-time Defense Strategy using Spectral Projections [12.807619042576018]
  We propose a novel test-time defense strategy called Robust Feature Inference (RFI) RFI is easy to integrate with any existing (robust) training procedure without additional test-time computation. We show that RFI improves robustness across adaptive and transfer attacks consistently.
  arXiv  Detail & Related papers  (2023-07-21T16:18:58Z)
• Blending adversarial training and representation-conditional purification via aggregation improves adversarial robustness [6.484231366444063]
  CARSO is able to defend itself against adaptive end-to-end white-box attacks devised for defences.<n>Our method improves by a significant margin the state-of-the-art for Cifar-10, Cifar-100, and TinyImageNet-200.
  arXiv  Detail & Related papers  (2023-05-25T09:04:31Z)
• Adaptive Feature Alignment for Adversarial Training [56.17654691470554]
  CNNs are typically vulnerable to adversarial attacks, which pose a threat to security-sensitive applications. We propose the adaptive feature alignment (AFA) to generate features of arbitrary attacking strengths. Our method is trained to automatically align features of arbitrary attacking strength.
  arXiv  Detail & Related papers  (2021-05-31T17:01:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.