分析100多个文生图模型，发现艺术类模型偏见更严重。
Exploring Bias in over 100 Text-to-Image Generative Models
- 从分布、幻觉、漏生成三维度评估模型偏见
- 艺术风格模型偏见显著，基础模型偏见逐渐降低
- 提供大规模评估数据集，助力负责任AI开发
我们研究了文本到图像生成模型随时间的偏见趋势，重点关注Hugging Face等开放平台带来的模型普及。尽管这些平台推动了AI民主化，但也加速了固有偏见模型的传播，这些模型常由特定任务微调形成。确保伦理透明的AI部署需依赖强大的评估框架和可量化的偏见指标。为此，我们从三个关键维度评估偏见：(i) 分布偏见，(ii) 生成幻觉，(iii) 生成漏失率。通过对超过100个模型的分析，揭示了偏见模式随时间与生成任务的变化规律。结果显示，艺术类及风格迁移模型存在显著偏见，而得益于更广泛训练分布的基础模型，其偏见正逐步减弱。本研究构建了大规模评估语料库，为偏见研究与缓解策略提供支持，促进更负责任的AI发展。
原文摘要 · Abstract (English)
We investigate bias trends in text-to-image generative models over time, focusing on the increasing availability of models through open platforms like Hugging Face. While these platforms democratize AI, they also facilitate the spread of inherently biased models, often shaped by task-specific fine-tuning. Ensuring ethical and transparent AI deployment requires robust evaluation frameworks and quantifiable bias metrics. To this end, we assess bias across three key dimensions: (i) distribution bias, (ii) generative hallucination, and (iii) generative miss-rate. Analyzing over 100 models, we reveal how bias patterns evolve over time and across generative tasks. Our findings indicate that artistic and style-transferred models exhibit significant bias, whereas foundation models, benefiting from broader training distributions, are becoming progressively less biased. By identifying these systemic trends, we contribute a large-scale evaluation corpus to inform bias research and mitigation strategies, fostering more responsible AI development. Keywords: Bias, Ethical AI, Text-to-Image, Generative Models, Open-Source Models
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。