mlsanityData to 7 Oct 2026

Finetuning Text-to-Image Diffusion Models for Fairness

arXiv 2311.07604 ↗First posted 11 Nov 2023cs.LG cs.AI cs.CVcited by 90

On the map: Diffusion models → Image generation, text-to-image and editing

Abstract (arXiv)

The rapid adoption of text-to-image diffusion models in society underscores an urgent need to address their biases. Without interventions, these biases could propagate a skewed worldview and restrict opportunities for minority groups. In this work, we frame fairness as a distributional alignment problem. Our solution consists of two main technical contributions: (1) a distributional alignment loss that steers specific characteristics of the generated images towards a user-defined target distribution, and (2) adjusted direct finetuning of diffusion model's sampling process (adjusted DFT), which leverages an adjusted gradient to directly optimize losses defined on the generated images. Empirically, our method markedly reduces gender, racial, and their intersectional biases for occupational prompts. Gender bias is significantly reduced even when finetuning just five soft tokens. Crucially, our method supports diverse perspectives of fairness beyond absolute equality, which is demonstrated by controlling age to a 75% young and 25% old distribution while simultaneously debiasing gender and race. Finally, our method is scalable: it can debias multiple concepts at once by simply including these prompts in the finetuning data. We share code and various fair diffusion model adaptors at https://sail-sg.github.io/finetune-fair-diffusion/.

Key facts. Cited by 90 papers in the map (data through 7 Oct 2026). Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models. Builds on 38 of its 52 references that are in the map. First posted on arXiv 11 Nov 2023.

Builds on

38 of 52 references.

Closest ancestors · 8

Same direction, newest first

  1. 2023Editing Implicit Assumptions in Text-to-Image Diffusion Models14 Mar
  2. 2023Adding Conditional Control to Text-to-Image Diffusion Models10 Feb
  3. 2022InstructPix2Pix: Learning to Follow Image Editing Instructions17 Nov
  4. 2022DiffusionDB: A Large-scale Prompt Gallery Dataset for Text-to-Image Generative Models26 Oct
  5. 2022Generative Visual Prompt: Unifying Distributional Control of Pre-Trained Generative Models14 Sep
  6. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding23 May
  7. 2022Hierarchical Text-Conditional Image Generation with CLIP Latents13 Apr
  8. 2021GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models20 Dec

Other references · 30

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  4. 2022Training language models to follow instructions with human feedbackcited by 15,712
  5. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  6. 2022Classifier-Free Diffusion Guidancecited by 5,717
  7. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
  8. 2021Prefix-Tuning: Optimizing Continuous Prompts for Generationcited by 3,878
  9. 2017Deep reinforcement learning from human preferencescited by 3,846
  10. 2021The Power of Scale for Parameter-Efficient Prompt Tuningcited by 3,769
  11. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  12. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,634
18 more
  1. 2022An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversioncited by 2,231
  2. 2022DPM-Solver++: Fast Solver for Guided Sampling of Diffusion Probabilistic Modelscited by 924
  3. 2016Training Deep Nets with Sublinear Memory Costcited by 825
  4. 2022Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Modelscited by 501
  5. 2023Directly Fine-Tuning Diffusion Models on Differentiable Rewardscited by 431
  6. 2023Unified Concept Editing in Diffusion Modelscited by 377
  7. 2023Emu: Enhancing Image Generation Models Using Photogenic Needles in a Haystackcited by 278
  8. 2022Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scalecited by 262
  9. 2023Efficient Diffusion Training via Min-SNR Weighting Strategycited by 242
  10. 2023Stable Bias: Analyzing Societal Representations in Diffusion Modelscited by 225
  11. 2023Fair Diffusion: Instructing Text-to-Image Generation Models on Fairnesscited by 149
  12. 2023Debiasing Vision-Language Models via Biased Promptscited by 147
  13. 2023End-to-End Diffusion Latent Optimization Improves Classifier Guidancecited by 107
  14. 2022How well can Text-to-Image Generative Models understand Ethical Natural Language Interventions?cited by 92
  15. 2023The Bias Amplification Paradox in Text-to-Image Generationcited by 69
  16. 2022Contrastive Language-Vision AI Models Pretrained on Web-Scraped Multimodal Data Exhibit Sexual Objectification Biascited by 43
  17. 2021SphereFace2: Binary Classification is All You Need for Deep Face Recognitioncited by 42
  18. 2023T2IAT: Measuring Valence and Stereotypical Biases in Text-to-Image Generationcited by 39

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 1; 2024: 37; 2025: 31; 2026: 21.

90 papers.

Sort:

By area:

  1. 2024Improved Distribution Matching Distillation for Fast Image Synthesiscited by 648
  2. 2024Survey of Bias In Text-to-Image Generation: Definition, Evaluation, and Mitigationcited by 66
  3. 2025T2ISafety: Benchmark for Assessing Fairness, Toxicity, and Privacy in Image Generationcited by 47
  4. 2024Margin-aware Preference Optimization for Aligning Diffusion Models without Referencecited by 45
  5. 2024VBench++: Comprehensive and Versatile Benchmark Suite for Video Generative Modelscited by 42
  6. 2024Token Merging for Training-Free Semantic Binding in Text-to-Image Synthesiscited by 35
  7. 2024Alignment of Diffusion Models: Fundamentals, Challenges, and Futurecited by 34
  8. 2023Fair Text-to-Image Diffusion via Fair Mappingcited by 31
  9. 2024Conditional Image Synthesis with Diffusion Models: A Surveycited by 29
  10. 2024Would Deep Generative Models Amplify Bias in Future Models?cited by 24
  11. 2024Rethinking Training for De-biasing Text-to-Image Generation: Unlocking the Potential of Stable Diffusioncited by 22
  12. 2024Shielded Diffusion: Generating Novel and Diverse Images using Sparse Repellencycited by 21
38 more
  1. 2024Constrained Diffusion Models via Dual Trainingcited by 19
  2. 2024BIGbench: A Unified Benchmark for Evaluating Multi-dimensional Social Biases in Text-to-Image Modelscited by 17
  3. 2024BendVLM: Test-Time Debiasing of Vision-Language Embeddingscited by 17
  4. 2025Kernel-based Unsupervised Embedding Alignment for Enhanced Visual Representation in Vision-language Modelscited by 12
  5. 2024Replication in Visual Diffusion Models: A Survey and Outlookcited by 11
  6. 2024MIST: Mitigating Intersectional Bias with Disentangled Cross-Attention Editing in Text-to-Image Diffusion Modelscited by 10
  7. 2025Fair Generation without Unfair Distortions: Debiasing Text-to-Image Generation with Entanglement-Free Attentioncited by 9
  8. 2024FaceScore: Benchmarking and Enhancing Face Quality in Human Generationcited by 8
  9. 2024Thinking Racial Bias in Fair Forgery Detection: Models, Datasets and Evaluationscited by 8
  10. 2024FairGen: Enhancing Fairness in Text-to-Image Diffusion Models via Self-Discovering Latent Directionscited by 8
  11. 2025FairGen: Controlling Sensitive Attributes for Fair Generations in Diffusion Models via Adaptive Latent Guidancecited by 7
  12. 2025Bias Is a Subspace, Not a Coordinate: A Geometric Rethinking of Post-hoc Debiasing in Vision-Language Modelscited by 7
  13. 2024Trustworthy Text-to-Image Diffusion Models: A Timely and Focused Surveycited by 6
  14. 2025Can we Debias Social Stereotypes in AI-Generated Images? Examining Text-to-Image Outputs and User Perceptionscited by 6
  15. 2025FairImagen: Post-Processing for Bias Mitigation in Text-to-Image Modelscited by 6
  16. 2024How Fair is Your Diffusion Recommender Model?cited by 5
  17. 2025Debiasing Diffusion Model: Enhancing Fairness through Latent Representation Learning in Stable Diffusion Modelcited by 5
  18. 2025Calibrating Generative Models to Distributional Constraintscited by 5
  19. 2025BlurGuard: A Simple Approach for Robustifying Image Protection Against AI-Powered Editingcited by 5
  20. 2024Ethical-Lens: Curbing Malicious Usages of Open-Source Text-to-Image Modelscited by 4
  21. 2024InvDiff: Invariant Guidance for Bias Mitigation in Diffusion Modelscited by 4
  22. 2024FairDiffusion: Enhancing Equity in Latent Diffusion Models via Fair Bayesian Perturbationcited by 4
  23. 2025ADT: Tuning Diffusion Models with Adversarial Supervisioncited by 4
  24. 2025Deconstructing Bias: A Multifaceted Framework for Diagnosing Cultural and Compositional Inequities in Text-to-Image Generative Modelscited by 4
  25. 2024FairCoT: Enhancing Fairness in Text-to-Image Generation via Chain of Thought Reasoning with Multimodal Large Language Modelscited by 3
  26. 2025Multi-Group Proportional Representation for Text-to-Image Modelscited by 3
  27. 2025RespoDiff: Dual-Module Bottleneck Transformation for Responsible & Faithful T2I Generationcited by 3
  28. 2025Responsible Diffusion: A Comprehensive Survey on Safety, Ethics, and Trust in Diffusion Modelscited by 3
  29. 2025Video Text Preservation with Synthetic Text-Rich Videoscited by 3
  30. 2025Fully Unsupervised Self-debiasing of Text-to-Image Diffusion Modelscited by 3
  31. 2024FoolSDEdit: Deceptively Steering Your Edits Towards Targeted Attribute-aware Distributioncited by 2
  32. 2024Biasing & Debiasing based Approach Towards Fair Knowledge Transfer for Equitable Skin Analysiscited by 2
  33. 2024Perceptual Fairness in Image Restorationcited by 2
  34. 2024Fair Text to Medical Image Diffusion Model with Subgroup Distribution Aligned Tuningcited by 2
  35. 2024CounterFace: A Synthetic Face Dataset for Fine-Grained Counterfactual Evaluation of Face Recognition Systemscited by 2
  36. 2024GradBias: Unveiling Word Influence on Bias in Text-to-Image Generative Modelscited by 2
  37. 2024On the Fairness, Diversity and Reliability of Text-to-Image Generative Modelscited by 2
  38. 2025PoGDiff: Product-of-Gaussians Diffusion Models for Imbalanced Text-to-Image Generationcited by 2

50 most cited of 90 All 90 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024Bias Begets Bias: The Impact of Biased Embeddings on Diffusion Models
  2. 2023Leveraging Diffusion Perturbations for Measuring Fairness in Computer Vision
  3. 2023Exploring Social Bias in Downstream Applications of Text-to-Image Foundation Models
  4. 2025BiasMap: Leveraging Cross-Attentions to Discover and Mitigate Hidden Social Biases in Text-to-Image Generation
  5. 2025Prompting Away Stereotypes? Evaluating Bias in Text-to-Image Models for Occupations
  6. 2024Generated Bias: Auditing Internal Bias Dynamics of Text-To-Image Generative Models
  7. 2025Hidden Bias in the Machine: Stereotypes in Text-to-Image Models
  8. 2025A Framework for Benchmarking Fairness-Utility Trade-offs in Text-to-Image Models via Pareto Frontiers