Selective Test-Time Debiasing for CLIP via Reward Gating

• URL: http://arxiv.org/abs/2607.00423v1
• Date: Wed, 01 Jul 2026 04:33:26 GMT
• Title: Selective Test-Time Debiasing for CLIP via Reward Gating
• Abstract summary: Vision language models (VLMs) demonstrate strong zero-shot performance, but often perpetuate social stereotypes in person-centric queries.<n>We introduce Reward-Gated Test-Time Adaptation (RG-TTA), a reinforcement learning-based test-time adaptation framework.<n>RG-TTA adaptively triggers fairness regularization based on the bias sensitivity of each input during test-time policy adaptation.
• Score: 9.837249813808603
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Vision language models (VLMs) demonstrate strong zero-shot performance, but often perpetuate social stereotypes in person-centric queries, yielding skewed demographic distributions. Current debiasing methods apply uniform bias corrections across all input queries regardless of their bias sensitivity, creating a fundamental fairness--utility trade-off. Strong debiasing distorts semantically meaningful information in bias-insensitive queries, while weak debiasing fails to mitigate stereotypes in bias-sensitive ones. This one-size-fits-all approach hampers simultaneously achieving high utility on bias-insensitive queries and fairness on bias-sensitive queries. We introduce Reward-Gated Test-Time Adaptation (RG-TTA), a reinforcement learning-based test-time adaptation framework that selectively applies debiasing based on input sensitivity. RG-TTA adaptively triggers fairness regularization based on the bias sensitivity of each input during test-time policy adaptation, while focusing exclusively on optimizing cross-modal alignment for bias-insensitive inputs. Experiments on fairness benchmarks (e.g., FairFace, UTKFace) demonstrate substantial bias reduction while simultaneously improving zero-shot utility, resolving the trade-off of uniform debiasing.

Related papers

• Making Bias Non-Predictive: Training Robust LLM Judges via Reinforcement Learning [91.8584139564909]
  Large language models (LLMs) increasingly serve as automated judges, yet they remain susceptible to cognitive biases.<n>We propose Epistemic Independence Training (EIT), a reinforcement learning framework grounded in a key principle.<n>EIT operationalizes this through a balanced conflict strategy where bias signals are equally likely to support correct and incorrect answers.
  arXiv  Detail & Related papers  (2026-02-02T01:43:48Z)
• BLADE: Bias-Linked Adaptive DEbiasing [2.7352017408152083]
  BLADE is a generative debiasing framework that requires no prior knowledge of bias or bias-conflicting samples.<n>We evaluate BLADE on multiple benchmark datasets and show that it significantly outperforms state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-10-05T12:28:54Z)
• Identifying and Mitigating Social Bias Knowledge in Language Models [52.52955281662332]
  We propose a novel debiasing approach, Fairness Stamp (FAST), which enables fine-grained calibration of individual social biases.<n>FAST surpasses state-of-the-art baselines with superior debiasing performance.<n>This highlights the potential of fine-grained debiasing strategies to achieve fairness in large language models.
  arXiv  Detail & Related papers  (2024-08-07T17:14:58Z)
• Going Beyond Popularity and Positivity Bias: Correcting for Multifactorial Bias in Recommender Systems [74.47680026838128]
  Two typical forms of bias in user interaction data with recommender systems (RSs) are popularity bias and positivity bias. We consider multifactorial selection bias affected by both item and rating value factors. We propose smoothing and alternating gradient descent techniques to reduce variance and improve the robustness of its optimization.
  arXiv  Detail & Related papers  (2024-04-29T12:18:21Z)
• Take Care of Your Prompt Bias! Investigating and Mitigating Prompt Bias in Factual Knowledge Extraction [56.17020601803071]
  Recent research shows that pre-trained language models (PLMs) suffer from "prompt bias" in factual knowledge extraction. This paper aims to improve the reliability of existing benchmarks by thoroughly investigating and mitigating prompt bias.
  arXiv  Detail & Related papers  (2024-03-15T02:04:35Z)
• Bias-Conflict Sample Synthesis and Adversarial Removal Debias Strategy for Temporal Sentence Grounding in Video [67.24316233946381]
  Temporal Sentence Grounding in Video (TSGV) is troubled by dataset bias issue. We propose the bias-conflict sample synthesis and adversarial removal debias strategy (BSSARD)
  arXiv  Detail & Related papers  (2024-01-15T09:59:43Z)
• On Comparing Fair Classifiers under Data Bias [42.43344286660331]
  We study the effect of varying data biases on the accuracy and fairness of fair classifiers. Our experiments show how to integrate a measure of data bias risk in the existing fairness dashboards for real-world deployments.
  arXiv  Detail & Related papers  (2023-02-12T13:04:46Z)
• Self-supervised debiasing using low rank regularization [59.84695042540525]
  Spurious correlations can cause strong biases in deep neural networks, impairing generalization ability. We propose a self-supervised debiasing framework potentially compatible with unlabeled samples. Remarkably, the proposed debiasing framework significantly improves the generalization performance of self-supervised learning baselines.
  arXiv  Detail & Related papers  (2022-10-11T08:26:19Z)
• BiaSwap: Removing dataset bias with bias-tailored swapping augmentation [20.149645246997668]
  Deep neural networks often make decisions based on the spurious correlations inherent in the dataset, failing to generalize in an unbiased data distribution. This paper proposes a novel bias-tailored augmentation-based approach, BiaSwap, for learning debiased representation without requiring supervision on the bias type.
  arXiv  Detail & Related papers  (2021-08-23T08:35:26Z)
• Bias-Tolerant Fair Classification [20.973916494320246]
  label bias and selection bias are two reasons in data that will hinder the fairness of machine-learning outcomes. We propose a Bias-TolerantFAirRegularizedLoss (B-FARL) which tries to regain the benefits using data affected by label bias and selection bias. B-FARL takes the biased data as input, calls a model that approximates the one trained with fair but latent data, and thus prevents discrimination without constraints required.
  arXiv  Detail & Related papers  (2021-07-07T13:31:38Z)

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.