mlsanityData to 7 Oct 2026

Analyzing Bias in Diffusion-based Face Generation Models

arXiv 2305.06402 ↗First posted 10 May 2023cs.CVcited by 45

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

Abstract (arXiv)

Diffusion models are becoming increasingly popular in synthetic data generation and image editing applications. However, these models can amplify existing biases and propagate them to downstream applications. Therefore, it is crucial to understand the sources of bias in their outputs. In this paper, we investigate the presence of bias in diffusion-based face generation models with respect to attributes such as gender, race, and age. Moreover, we examine how dataset size affects the attribute composition and perceptual quality of both diffusion and Generative Adversarial Network (GAN) based face generation models across various attribute classes. Our findings suggest that diffusion models tend to worsen distribution bias in the training data for various attributes, which is heavily influenced by the size of the dataset. Conversely, GAN models trained on balanced datasets with a larger number of samples show less bias across different attributes.

Key facts. Cited by 45 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 27 of its 40 references that are in the map. First posted on arXiv 10 May 2023.

Builds on

27 of 40 references.

Closest ancestors · 2

Same direction, newest first

  1. 2023Synthetic Data from Diffusion Models Improves ImageNet Classification17 Apr
  2. 2022Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models7 Dec

Other references · 25

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  4. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
  5. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  6. 2021Image Super-Resolution via Iterative Refinementcited by 1,506
  7. 2020Training Generative Adversarial Networks with Limited Datacited by 1,184
  8. 2021Cascaded Diffusion Models for High Fidelity Image Generationcited by 1,155
  9. 2023Extracting Training Data from Diffusion Modelscited by 723
  10. 2022Perception Prioritized Training of Diffusion Modelscited by 301
  11. 2023Stable Bias: Analyzing Societal Representations in Diffusion Modelscited by 225
  12. 2019Mitigate Bias in Face Recognition using Skewness-Aware Reinforcement Learningcited by 145
13 more
  1. 2023Social Biases through the Text-to-Image Generation Lenscited by 126
  2. 2020Generative Adversarial Networks for Image and Video Synthesis: Algorithms and Applicationscited by 48
  3. 2020Improving the Fairness of Deep Generative Models without Retrainingcited by 46
  4. 2020Inclusive GAN: Improving Data and Minority Coverage in Generative Modelscited by 44
  5. 2019Characterizing the Variability in Face Recognition Accuracy Relative to Racecited by 40
  6. 2021PASS: Protected Attribute Suppression System for Mitigating Bias in Face Recognitioncited by 38
  7. 2023Synthetic Data for Face Recognition: Current State and Future Prospectscited by 38
  8. 2020How Does Gender Balance In Training Data Affect Face Recognition Accuracy?cited by 37
  9. 2022When do GANs replicate? On the choice of dataset sizecited by 34
  10. 2021Gendered Differences in Face Recognition Accuracy Explained by Hairstyles, Makeup, and Facial Morphologycited by 32
  11. 2022Studying Bias in GANs through the Lens of Racecited by 31
  12. 2022Face Recognition Accuracy Across Demographics: Shining a Light Into the Problemcited by 24
  13. 2022RepFair-GAN: Mitigating Representation Bias in GANs Using Gradient Clippingcited by 9

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 8; 2024: 18; 2025: 11; 2026: 8.

45 papers.

Sort:

By area:

  1. 2023On the Design Fundamentals of Diffusion Models: A Surveycited by 59
  2. 2024Balancing Act: Distribution-Guided Debiasing in Diffusion Modelscited by 50
  3. 2023Photoswap: Personalized Subject Swapping in Imagescited by 47
  4. 2023The Journey, Not the Destination: How Data Guides Diffusion Modelscited by 42
  5. 2024FairRAG: Fair Human Generation via Fair Retrieval Augmentationcited by 34
  6. 2024Rethinking Training for De-biasing Text-to-Image Generation: Unlocking the Potential of Stable Diffusioncited by 22
  7. 2024Harm Amplification in Text-to-Image Modelscited by 14
  8. 2024What do we learn from inverting CLIP models?cited by 13
  9. 2024Diffusion-driven GAN Inversion for Multi-Modal Face Image Generationcited by 12
  10. 2024Frame by Familiar Frame: Understanding Replication in Video Diffusion Modelscited by 11
  11. 2024Replication in Visual Diffusion Models: A Survey and Outlookcited by 11
  12. 2025Beyond Real Faces: Synthetic Datasets Can Achieve Reliable Recognition Performance without Privacy Compromisecited by 8
33 more
  1. 2023Exploring Social Bias in Downstream Applications of Text-to-Image Foundation Modelscited by 7
  2. 2024SDFD: Building a Versatile Synthetic Face Image Dataset with Diverse Attributescited by 6
  3. 2024VersusDebias: Universal Zero-Shot Debiasing for Text-to-Image Models via SLM-Based Prompt Engineering and Generative Adversarycited by 6
  4. 2025Debiasing Diffusion Model: Enhancing Fairness through Latent Representation Learning in Stable Diffusion Modelcited by 5
  5. 2025RL4Med-DDPO: Reinforcement Learning for Controlled Guidance Towards Diverse Medical Image Generation using Vision-Language Foundation Modelscited by 5
  6. 2023Limitations of Face Image Generationcited by 4
  7. 2023DiffSpectralNet : Unveiling the Potential of Diffusion Models for Hyperspectral Image Classificationcited by 4
  8. 2023Gaussian Harmony: Attaining Fairness in Diffusion-based Face Generation Modelscited by 4
  9. 2025Diffusing DeBias: Synthetic Bias Amplification for Model Debiasingcited by 3
  10. 2025RespoDiff: Dual-Module Bottleneck Transformation for Responsible & Faithful T2I Generationcited by 3
  11. 2025Fully Unsupervised Self-debiasing of Text-to-Image Diffusion Modelscited by 3
  12. 2024Diverse Feature Learning by Self-distillation and Resetcited by 2
  13. 2024Tuning-Free Visual Customization via View Iterative Self-Attention Controlcited by 2
  14. 2024The Fabrication of Reality and Fantasy: Scene Generation with LLM-Assisted Prompt Interpretationcited by 2
  15. 2024Fairer Analysis and Demographically Balanced Face Generation for Fairer Face Verificationcited by 2
  16. 2023What's color got to do with it? Face recognition in grayscalecited by 1
  17. 2024Mitigating the Impact of Attribute Editing on Face Recognitioncited by 1
  18. 2024Bias Begets Bias: The Impact of Biased Embeddings on Diffusion Modelscited by 1
  19. 2025Using LLMs as prompt modifier to avoid biases in AI image generatorscited by 1
  20. 2025Palette Aligned Image Diffusioncited by 1
  21. 2025From Symptoms to Systems: An Expert-Guided Approach to Understanding Risks of Generative AI for Eating Disorderscited by 1
  22. 2026Discount Model Search for Quality Diversity Optimization in High-Dimensional Measure Spacescited by 1
  23. 2026Training-Free Debiasing of Diffusion Models via CLIP-Guided Denoising Optimizationcited by 1
  24. 2024Toward Fairer Face Recognition Datasetscited by 0
  25. 2024ContRail: A Framework for Realistic Railway Image Synthesis using ControlNetcited by 0
  26. 2025Generative AI in Collaborative Academic Report Writing: Advantages, Disadvantages, and Ethical Considerationscited by 0
  27. 2025Bias Analysis in Unconditional Image Generative Modelscited by 0
  28. 2026RAIGen: Rare Attribute Identification in Text-to-Image Generative Modelscited by 0
  29. 2026Operationalizing Fairness in Text-to-Image Models: A Survey of Bias, Fairness Audits and Mitigation Strategiescited by 0
  30. 2026Efficient bias mitigation in T2I diffusion models using Concept Graphscited by 0
  31. 2026Learning Late, Guiding Early: Timestep-Decoupled Semantic Guidance for Fair Face Generationcited by 0
  32. 2026A Large-scale Evaluation of Text-guided Models for Facial Editingcited by 0
  33. 2026Debias Anything: Fairness with Diversity without Supervision in Diffusion Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Uncovering Bias in Face Generation Models
  2. 2023Conditioning Diffusion Models via Attributes and Semantic Masks for Face Generation
  3. 2025Deeper Diffusion Models Amplify Bias
  4. 2024Would Deep Generative Models Amplify Bias in Future Models?
  5. 2024AI-generated faces influence gender stereotypes and racial homogenization
  6. 2024Generated Bias: Auditing Internal Bias Dynamics of Text-To-Image Generative Models
  7. 2024Towards the Detection of AI-Synthesized Human Face Images
  8. 2024LDFaceNet: Latent Diffusion-based Network for High-Fidelity Deepfake Generation