Freeze and Reveal: Exposing Modality Bias in Vision-Language Models

• URL: http://arxiv.org/abs/2508.07432v1
• Date: Sun, 10 Aug 2025 17:08:10 GMT
• Title: Freeze and Reveal: Exposing Modality Bias in Vision-Language Models
• Authors: Vivek Hruday Kavuri, Vysishtya Karanam, Venkata Jahnavi Venkamsetty, Kriti Madumadukala, Lakshmipathi Balaji Darur, Ponnurangam Kumaraguru,
• Abstract summary: Vision Language Models achieve impressive multi-modal performance but often inherit gender biases from their training data.<n>In this work, we dissect the contributions of vision and text backbones to these biases by applying targeted debiasing.<n>Our work enables more targeted and effective bias mitigation strategies in future multi-modal systems.
• Score: 5.368169150936655
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Vision Language Models achieve impressive multi-modal performance but often inherit gender biases from their training data. This bias might be coming from both the vision and text modalities. In this work, we dissect the contributions of vision and text backbones to these biases by applying targeted debiasing using Counterfactual Data Augmentation and Task Vector methods. Inspired by data-efficient approaches in hate-speech classification, we introduce a novel metric, Degree of Stereotypicality and a corresponding debiasing method, Data Augmentation Using Degree of Stereotypicality - DAUDoS, to reduce bias with minimal computational cost. We curate a gender annotated dataset and evaluate all methods on VisoGender benchmark to quantify improvements and identify dominant source of bias. Our results show that CDA reduces the gender gap by 6% and DAUDoS by 3% but using only one-third of the data. Both methods also improve the model's ability to correctly identify gender in images by 3%, with DAUDoS achieving this improvement using only almost one-third of training data. From our experiment's, we observed that CLIP's vision encoder is more biased whereas PaliGemma2's text encoder is more biased. By identifying whether bias stems more from vision or text encoders, our work enables more targeted and effective bias mitigation strategies in future multi-modal systems.

Related papers

• Exploring and Mitigating Gender Bias in Encoder-Based Transformer Models [0.0]
  This paper investigates gender bias in contextualized word embeddings, a crucial component of transformer-based models.<n>To quantify the degree of bias, we introduce a novel metric, MALoR, which assesses bias based on model probabilities for filling masked tokens.<n>Our experiments reveal significant reductions in gender bias scores across different pronoun pairs.
  arXiv  Detail & Related papers  (2025-11-01T11:49:44Z)
• Data Matters Most: Auditing Social Bias in Contrastive Vision Language Models [28.944990804599893]
  We disentangle three design factors -- model size, training-data scale, and training-data source -- by comparing CLIP and OpenCLIP.<n>We also evaluate three post-hoc, test-time debiasing strategies -- Bias Prompts, Prompt Array, and SANER.
  arXiv  Detail & Related papers  (2025-01-22T21:08:30Z)
• Revealing and Reducing Gender Biases in Vision and Language Assistants (VLAs) [82.57490175399693]
  We study gender bias in 22 popular image-to-text vision-language assistants (VLAs)<n>Our results show that VLAs replicate human biases likely present in the data, such as real-world occupational imbalances.<n>To eliminate the gender bias in these models, we find that fine-tuning-based debiasing methods achieve the best trade-off between debiasing and retaining performance.
  arXiv  Detail & Related papers  (2024-10-25T05:59:44Z)
• debiaSAE: Benchmarking and Mitigating Vision-Language Model Bias [1.3995965887921709]
  We analyze demographic biases across five models and six datasets.<n>Portrait datasets like UTKFace and CelebA are the best tools for bias detection.<n>Our debiasing method improves fairness, gaining 5-15 points in performance over the baseline.
  arXiv  Detail & Related papers  (2024-10-17T02:03:27Z)
• Images Speak Louder than Words: Understanding and Mitigating Bias in Vision-Language Model from a Causal Mediation Perspective [13.486497323758226]
  Vision-language models pre-trained on extensive datasets can inadvertently learn biases by correlating gender information with objects or scenarios.<n>We propose a framework that incorporates causal mediation analysis to measure and map the pathways of bias generation and propagation.
  arXiv  Detail & Related papers  (2024-07-03T05:19:45Z)
• GenderBias-\emph{VL}: Benchmarking Gender Bias in Vision Language Models via Counterfactual Probing [72.0343083866144]
  This paper introduces the GenderBias-emphVL benchmark to evaluate occupation-related gender bias in Large Vision-Language Models. Using our benchmark, we extensively evaluate 15 commonly used open-source LVLMs and state-of-the-art commercial APIs. Our findings reveal widespread gender biases in existing LVLMs.
  arXiv  Detail & Related papers  (2024-06-30T05:55:15Z)
• The Impact of Debiasing on the Performance of Language Models in Downstream Tasks is Underestimated [70.23064111640132]
  We compare the impact of debiasing on performance across multiple downstream tasks using a wide-range of benchmark datasets. Experiments show that the effects of debiasing are consistently emphunderestimated across all tasks.
  arXiv  Detail & Related papers  (2023-09-16T20:25:34Z)
• VisoGender: A dataset for benchmarking gender bias in image-text pronoun resolution [80.57383975987676]
  VisoGender is a novel dataset for benchmarking gender bias in vision-language models. We focus on occupation-related biases within a hegemonic system of binary gender, inspired by Winograd and Winogender schemas. We benchmark several state-of-the-art vision-language models and find that they demonstrate bias in resolving binary gender in complex scenes.
  arXiv  Detail & Related papers  (2023-06-21T17:59:51Z)
• Balancing the Picture: Debiasing Vision-Language Datasets with Synthetic Contrast Sets [52.77024349608834]
  Vision-language models can perpetuate and amplify societal biases learned during pre-training on uncurated image-text pairs from the internet. COCO Captions is the most commonly used dataset for evaluating bias between background context and the gender of people in-situ. We propose a novel dataset debiasing pipeline to augment the COCO dataset with synthetic, gender-balanced contrast sets.
  arXiv  Detail & Related papers  (2023-05-24T17:59:18Z)
• D-BIAS: A Causality-Based Human-in-the-Loop System for Tackling Algorithmic Bias [57.87117733071416]
  We propose D-BIAS, a visual interactive tool that embodies human-in-the-loop AI approach for auditing and mitigating social biases. A user can detect the presence of bias against a group by identifying unfair causal relationships in the causal network. For each interaction, say weakening/deleting a biased causal edge, the system uses a novel method to simulate a new (debiased) dataset.
  arXiv  Detail & Related papers  (2022-08-10T03:41:48Z)
• An Empirical Survey of the Effectiveness of Debiasing Techniques for Pre-Trained Language Models [4.937002982255573]
  Recent work has shown that pre-trained language models capture social biases from the text corpora they are trained on. Five recently proposed debiasing techniques: Counterfactual Data Augmentation, Dropout, Iterative Nullspace Projection, Self-Debias, and SentenceDebias. We quantify the effectiveness of each technique using three different bias benchmarks while also measuring the impact of these techniques on a model's language modeling ability.
  arXiv  Detail & Related papers  (2021-10-16T09:40:30Z)
• Mitigating Gender Bias in Captioning Systems [56.25457065032423]
  Most captioning models learn gender bias, leading to high gender prediction errors, especially for women. We propose a new Guided Attention Image Captioning model (GAIC) which provides self-guidance on visual attention to encourage the model to capture correct gender visual evidence.
  arXiv  Detail & Related papers  (2020-06-15T12:16:19Z)

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.