Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
41 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

FAIntbench: A Holistic and Precise Benchmark for Bias Evaluation in Text-to-Image Models (2405.17814v5)

Published 28 May 2024 in cs.CV and cs.AI

Abstract: The rapid development and reduced barriers to entry for Text-to-Image (T2I) models have raised concerns about the biases in their outputs, but existing research lacks a holistic definition and evaluation framework of biases, limiting the enhancement of debiasing techniques. To address this issue, we introduce FAIntbench, a holistic and precise benchmark for biases in T2I models. In contrast to existing benchmarks that evaluate bias in limited aspects, FAIntbench evaluate biases from four dimensions: manifestation of bias, visibility of bias, acquired attributes, and protected attributes. We applied FAIntbench to evaluate seven recent large-scale T2I models and conducted human evaluation, whose results demonstrated the effectiveness of FAIntbench in identifying various biases. Our study also revealed new research questions about biases, including the side-effect of distillation. The findings presented here are preliminary, highlighting the potential of FAIntbench to advance future research aimed at mitigating the biases in T2I models. Our benchmark is publicly available to ensure the reproducibility.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (45)
  1. Y. Anagnostou. “white ethnicity”: A reappraisal. Italian American Review, 3(2):99–128, 2013.
  2. Hrs-bench: Holistic, reliable and scalable benchmark for text-to-image models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 20041–20053, 2023.
  3. How well can text-to-image generative models understand ethical natural language interventions? arXiv preprint arXiv:2210.15230, 2022.
  4. S. Benthall and B. D. Haynes. Racial categories in machine learning. In Proceedings of the conference on fairness, accountability, and transparency, pages 289–298, 2019.
  5. Improving image generation with better captions. Computer Science. https://cdn. openai. com/papers/dall-e-3. pdf, 2(3):8, 2023.
  6. Pixart-α𝛼\alphaitalic_α: Fast training of diffusion transformer for photorealistic text-to-image synthesis. arXiv preprint arXiv:2310.00426, 2023.
  7. Pixart-σ𝜎\sigmaitalic_σ: Weak-to-strong training of diffusion transformer for 4k text-to-image generation. arXiv preprint arXiv:2403.04692, 2024.
  8. Tibet: Identifying and evaluating biases in text-to-image generative models. arXiv preprint arXiv:2312.01261, 2023.
  9. Dall-eval: Probing the reasoning skills and social biases of text-to-image generation models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3043–3054, 2023.
  10. A. Chouldechova. Fair prediction with disparate impact: A study of bias in recidivism prediction instruments. Big data, 5(2):153–163, 2017.
  11. Precisedebias: An automatic prompt engineering approach for generative ai to mitigate image demographic biases. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 8596–8605, 2024.
  12. Cogview2: Faster and better text-to-image generation via hierarchical transformers. Advances in Neural Information Processing Systems, 35:16890–16902, 2022.
  13. Scaling rectified flow transformers for high-resolution image synthesis. arXiv preprint arXiv:2403.03206, 2024.
  14. E. Ferrara. Fairness and bias in artificial intelligence: A brief survey of sources, impacts, and mitigation strategies. Sci, 6(1):3, 2023.
  15. L. Fridell. This is not your grandparents’ prejudice: The implications of the modern science of bias for police training. Translational Criminology, 5(1):10–11, 2013.
  16. Fair diffusion: Instructing text-to-image generation models on fairness. arXiv preprint arXiv:2302.10893, 2023.
  17. Bias and fairness in large language models: A survey. arXiv preprint arXiv:2309.00770, 2023.
  18. Unified concept editing in diffusion models. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 5111–5120, 2024.
  19. Datasheets for datasets. Communications of the ACM, 64(12):86–92, 2021.
  20. A review of possible effects of cognitive biases on interpretation of rule-based machine learning models. Artificial Intelligence, 295:103458, 2021.
  21. Bias and ignorance in demographic perception. Psychonomic bulletin & review, 25:1606–1618, 2018.
  22. S. Larsson and F. Heintz. Transparency in artificial intelligence. Internet Policy Review, 9(2), 2020.
  23. Playground v2. 5: Three insights towards enhancing aesthetic quality in text-to-image generation. arXiv preprint arXiv:2402.17245, 2024.
  24. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. In International conference on machine learning, pages 19730–19742. PMLR, 2023.
  25. Sdxl-lightning: Progressive adversarial diffusion distillation. arXiv preprint arXiv:2402.13929, 2024.
  26. Stable bias: Analyzing societal representations in diffusion models. arXiv preprint arXiv:2303.11408, 2023.
  27. Latent consistency models: Synthesizing high-resolution images with few-step inference. arXiv preprint arXiv:2310.04378, 2023a.
  28. Lcm-lora: A universal stable-diffusion acceleration module. arXiv preprint arXiv:2311.05556, 2023b.
  29. A survey on bias and fairness in machine learning. ACM computing surveys (CSUR), 54(6):1–35, 2021.
  30. On distillation of guided diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14297–14306, 2023.
  31. J. Moule. Understanding unconscious bias and unintentional racism. Phi Delta Kappan, 90(5):320–326, 2009.
  32. Würstchen: An efficient architecture for large-scale text-to-image diffusion models. In The Twelfth International Conference on Learning Representations, 2023.
  33. Sdxl: Improving latent diffusion models for high-resolution image synthesis. arXiv preprint arXiv:2307.01952, 2023.
  34. Learning transferable visual models from natural language supervision. In International conference on machine learning, pages 8748–8763. PMLR, 2021.
  35. Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125, 1(2):3, 2022.
  36. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022.
  37. Adversarial diffusion distillation. arXiv preprint arXiv:2311.17042, 2023.
  38. Safe latent diffusion: Mitigating inappropriate degeneration in diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 22522–22531, 2023.
  39. Sdxs: Real-time one-step latent diffusion models with image conditions. arXiv preprint arXiv:2403.16627, 2024.
  40. World population prospects 2022: Summary of results. Technical Report UN DESA/POP/2022/TR/NO. 3, United Nations, 2022. URL https://population.un.org/wpp/.
  41. t. U.S. Bureau of Labor Statistics. Employed persons by detailed occupation and age : U.S. Bureau of Labor Statistics — bls.gov. https://www.bls.gov/cps/cpsaat11b.htm, 2023. [Accessed 12-05-2024].
  42. t. U.S. Census Bureau. Full-Time, Year-Round Workers & Median Earnings by Sex & Occupation — census.gov. https://www.census.gov/data/tables/time-series/demo/industry-occupation/median-earnings.html, 2022. [Accessed 12-05-2024].
  43. t. U.S. Department of Justice. Understanding bias: A resource guide. https://www.justice.gov/crs/file/1188566/dl?inline=. [Accessed 12-05-2024].
  44. D. Varona and J. L. Suárez. Discrimination, bias, fairness, and trustworthy ai. Applied Sciences, 12(12):5826, 2022.
  45. Survey of bias in text-to-image generation: Definition, evaluation, and mitigation. arXiv preprint arXiv:2404.01030, 2024.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (4)
  1. Hanjun Luo (8 papers)
  2. Ziye Deng (3 papers)
  3. Ruizhe Chen (32 papers)
  4. Zuozhu Liu (78 papers)
Citations (5)
View on Semantic Scholar
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets