mlsanityData to 7 Oct 2026

A Prompt Array Keeps the Bias Away: Debiasing Vision-Language Models with Adversarial Learning

arXiv 2203.11933 ↗First posted 22 Mar 2022cs.LG cs.CL cs.CVcited by 104

On the map: Multimodal LLMs / vision-language models → Social bias and fairness in VLMs

Abstract (arXiv)

Vision-language models can encode societal biases and stereotypes, but there are challenges to measuring and mitigating these multimodal harms due to lacking measurement robustness and feature degradation. To address these challenges, we investigate bias measures and apply ranking metrics for image-text representations. We then investigate debiasing methods and show that prepending learned embeddings to text queries that are jointly trained with adversarial debiasing and a contrastive loss reduces various bias measures with minimal degradation to the image-text representation.

Key facts. Cited by 104 papers in the map (data through 7 Oct 2026). Placed in Social bias and fairness in VLMs, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 62 of its 86 references that are in the map. First posted on arXiv 22 Mar 2022.

Builds on

62 of 86 references.

References · 62

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  4. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  5. 2020Language Models are Few-Shot Learnerscited by 32,199
  6. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  7. 2022Training language models to follow instructions with human feedbackcited by 15,712
  8. 2021Learning to Prompt for Vision-Language Modelscited by 2,269
  9. 2021Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processingcited by 2,222
  10. 2019A Survey on Bias and Fairness in Machine Learningcited by 1,911
  11. 2021Align before Fuse: Vision and Language Representation Learning with Momentum Distillationcited by 1,612
  12. 2022Conditional Prompt Learning for Vision-Language Modelscited by 1,427
38 more
  1. 2018Datasheets for Datasetscited by 1,371
  2. 2016Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddingscited by 1,304
  3. 2021Conceptual 12M: Pushing Web-Scale Image-Text Pre-Training To Recognize Long-Tail Visual Conceptscited by 1,209
  4. 2021Frozen in Time: A Joint Video and Image Encoder for End-to-End Retrievalcited by 1,175
  5. 2019A Comprehensive Survey on Transfer Learningcited by 1,154
  6. 2021Ethical and social risks of harm from Language Modelscited by 957
  7. 2019HowTo100M: Learning a Text-Video Embedding by Watching Hundred Million Narrated Video Clipscited by 946
  8. 2016Semantics derived automatically from language corpora contain human-like biasescited by 894
  9. 2020VGGSound: A Large-scale Audio-Visual Datasetcited by 641
  10. 2021LiT: Zero-Shot Transfer with Locked-image text Tuningcited by 572
  11. 2021SLIP: Self-supervision meets Language-Image Pre-trainingcited by 434
  12. 2017Age Progression/Regression by Conditional Adversarial Autoencodercited by 426
  13. 2017Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraintscited by 420
  14. 2017Algorithmic decision making and the cost of fairnesscited by 402
  15. 2019On Measuring Social Biases in Sentence Encoderscited by 397
  16. 2021Multimodal datasets: misogyny, pornography, and malignant stereotypescited by 326
  17. 2021ActionCLIP: A New Paradigm for Video Action Recognitioncited by 317
  18. 2021Prompting Visual-Language Models for Efficient Video Understandingcited by 309
  19. 2021WIT: Wikipedia-based Image Text Dataset for Multimodal Multilingual Machine Learningcited by 285
  20. 2019Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Themcited by 268
  21. 2017Word Embeddings Quantify 100 Years of Gender and Ethnic Stereotypescited by 251
  22. 2018Women also Snowboard: Overcoming Bias in Captioning Modelscited by 236
  23. 2015YFCC100M: The New Data in Multimedia Researchcited by 234
  24. 2019Gender Bias in Contextualized Word Embeddingscited by 203
  25. 2015Censoring Representations with an Adversarycited by 191
  26. 2020Detecting Emergent Intersectional Biases: Contextualized Word Embeddings Contain a Distribution of Human-like Biasescited by 167
  27. 2018Adversarial Removal of Demographic Attributes from Text Datacited by 166
  28. 2017No Classification without Representation: Assessing Geodiversity Issues in Open Data Sets for the Developing Worldcited by 152
  29. 2019Does Object Recognition Work for Everyone?cited by 149
  30. 2019Black is to Criminal as Caucasian is to Police: Detecting and Removing Multiclass Bias in Word Embeddingscited by 146
  31. 2021Bias Out-of-the-Box: An Empirical Analysis of Intersectional Occupational Biases in Popular Generative Language Modelscited by 141
  32. 2019Fairness-Aware Ranking in Search & Recommendation Systems with Application to LinkedIn Talent Searchcited by 137
  33. 2020To be Robust or to be Fair: Towards Fairness in Adversarial Trainingcited by 137
  34. 2020REVISE: A Tool for Measuring and Mitigating Bias in Visual Datasetscited by 136
  35. 2021Evaluating CLIP: Towards Characterization of Broader Capabilities and Downstream Implicationscited by 128
  36. 2021Uni-Perceiver: Pre-training Unified Architecture for Generic Perception for Zero-shot and Few-shot Taskscited by 124
  37. 2019Lessons from Archives: Strategies for Collecting Sociocultural Data in Machine Learningcited by 116
  38. 2018Turning a Blind Eye: Explicit Removal of Biases and Variation from Deep Neural Network Embeddingscited by 114

50 most cited of 62 All 62 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 8; 2023: 31; 2024: 30; 2025: 26; 2026: 9.

104 papers.

Sort:

By area:

  1. 2023Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Modelscited by 297
  2. 2023What does CLIP know about a red circle? Visual prompt engineering for VLMscited by 215
  3. 2023Fair Diffusion: Instructing Text-to-Image Generation Models on Fairnesscited by 149
  4. 2023Debiasing Vision-Language Models via Biased Promptscited by 147
  5. 2023DeAR: Debiasing Vision-Language Models with Additive Residualscited by 70
  6. 2022Contrastive Adapters for Foundation Model Group Robustnesscited by 62
  7. 2022A CLIP-Hitchhiker's Guide to Long Video Retrievalcited by 57
  8. 2024Examining Gender and Racial Bias in Large Vision-Language Models Using a Novel Dataset of Parallel Imagescited by 55
  9. 2023SocialCounterfactuals: Probing and Mitigating Intersectional Social Biases in Vision-Language Models with Counterfactual Examplescited by 53
  10. 2023VisoGender: A dataset for benchmarking gender bias in image-text pronoun resolutioncited by 52
  11. 2023Uncurated Image-Text Datasets: Shedding Light on Demographic Biascited by 51
  12. 2024Balancing Act: Distribution-Guided Debiasing in Diffusion Modelscited by 50
38 more
  1. 2023Towards Safe Self-Distillation of Internet-Scale Text-to-Image Diffusion Modelscited by 41
  2. 2023Exposing and Mitigating Spurious Correlations for Cross-Modal Retrievalcited by 35
  3. 2024CLIP the Bias: How Useful is Balancing Data in Multimodal Learning?cited by 35
  4. 2023Visual Tuningcited by 34
  5. 2023Fair Text-to-Image Diffusion via Fair Mappingcited by 31
  6. 2022Addressing Fairness Issues in Deep Learning-Based Medical Image Analysis: A Systematic Reviewcited by 27
  7. 2023Balancing the Picture: Debiasing Vision-Language Datasets with Synthetic Contrast Setscited by 27
  8. 2023Sociodemographic Bias in Language Models: A Survey and Forward Pathcited by 27
  9. 2023ToViLaG: Your Visual-Language Generative Model is Also An Evildoercited by 27
  10. 2024Identifying Implicit Social Biases in Vision-Language Modelscited by 27
  11. 2023Survey of Social Bias in Vision-Language Modelscited by 26
  12. 2023A Friendly Face: Do Text-to-Image Systems Rely on Stereotypes when the Input is Under-Specified?cited by 25
  13. 2023Stable Diffusion Exposed: Gender Bias from Prompt to Imagecited by 25
  14. 2024FairerCLIP: Debiasing CLIP's Zero-Shot Predictions using Functions in RKHSscited by 25
  15. 2023Zero-Shot Robustification of Zero-Shot Modelscited by 24
  16. 2022Debiasing Methods for Fairer Neural Models in Vision and Language Research: A Surveycited by 22
  17. 2023Consensus and Subjectivity of Skin Tone Annotation for ML Fairnesscited by 22
  18. 2022Looking for a Handsome Carpenter! Debiasing GPT-3 Job Advertisementscited by 21
  19. 2025Prompt-based Adaptation in Large-scale Vision Models: A Surveycited by 21
  20. 2022FairCLIP: Social Bias Elimination based on Attribute Prototype Learning and Representation Neutralizationcited by 20
  21. 2023Mitigating Test-Time Bias for Fair Image Retrievalcited by 20
  22. 2024SANER: Annotation-free Societal Attribute Neutralizer for Debiasing CLIPcited by 19
  23. 2024A Unified Debiasing Approach for Vision-Language Models across Modalities and Taskscited by 18
  24. 2023Model-Agnostic Gender Debiased Image Captioningcited by 17
  25. 2023Indian-BhED: A Dataset for Measuring India-Centric Biases in Large Language Modelscited by 17
  26. 2024BendVLM: Test-Time Debiasing of Vision-Language Embeddingscited by 17
  27. 2024Joint Vision-Language Social Bias Removal for CLIPcited by 17
  28. 2023Parrot Captions Teach CLIP to Spot Textcited by 16
  29. 2024Revealing and Reducing Gender Biases in Vision and Language Assistants (VLAs)cited by 15
  30. 2023Evaluation and Enhancement of Semantic Grounding in Large Vision-Language Modelscited by 13
  31. 2023Evaluating the Fairness of Discriminative Foundation Models in Computer Visioncited by 12
  32. 2024Fairness and Bias Mitigation in Computer Vision: A Surveycited by 12
  33. 2025A Large Scale Analysis of Gender Biases in Text-to-Image Generative Modelscited by 12
  34. 2024FairDeDup: Detecting and Mitigating Vision-Language Fairness Disparities in Semantic Dataset Deduplicationcited by 11
  35. 2024Social Perception of Faces in a Vision-Language Modelcited by 10
  36. 2025BBQ-V: Benchmarking Visual Stereotype Bias in Large Multimodal Modelscited by 10
  37. 2023Unpacking the Ethical Value Alignment in Big Modelscited by 9
  38. 2022Zero-shot Image Captioning by Anchor-augmented Vision-Language Space Alignmentcited by 8

50 most cited of 104 All 104 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026A Closed-Form Solution for Debiasing Vision-Language Models with Utility Guarantees Across Modalities and Tasks
  2. 2024debiaSAE: Benchmarking and Mitigating Vision-Language Model Bias
  3. 2025Freeze and Reveal: Exposing Modality Bias in Vision-Language Models
  4. 2026BiPrompt: Bilateral Prompt Optimization for Visual and Textual Debiasing in Vision-Language Models
  5. 2024Uncovering Bias in Large Vision-Language Models at Scale with Counterfactuals
  6. 2024Uncovering Bias in Large Vision-Language Models with Counterfactuals
  7. 2023MultiModal Bias: Introducing a Framework for Stereotypical Bias Assessment beyond Gender and Race in Vision Language Models
  8. 2024Images Speak Louder than Words: Understanding and Mitigating Bias in Vision-Language Model from a Causal Mediation Perspective