男性CEO与女性助理：双主体文本到图像生成中性别偏见的评估与缓解
The Male CEO and the Female Assistant: Evaluation and Mitigation of Gender Biases in Text-To-Image Generation of Dual Subjects
AI总结：
本文提出PST框架系统评估双主体T2I生成中的性别偏见，并设计FairCritic框架通过LLM评判模型检测偏见并自适应反馈，在PST上实现近乎完美的公平性。
摘要：
近期，DALLE-3等大规模文本到图像（T2I）模型在生成单人图像时已在减少性别刻板印象方面取得进展。然而，在生成包含多人的图像时，显著的偏见仍然存在。为了系统性地评估这一问题，我们提出了配对刻板印象测试（Paired Stereotype Test，PST）框架，该框架要求T2I模型分别描绘两个被赋予男性刻板印象社会身份和女性刻板印象社会身份的个体（例如“一位CEO”和“一位助理”）。这种对比设置往往会触发T2I模型生成带有性别刻板印象的图像。利用PST，我们评估了性别偏见的两个方面——众所周知的性别职业偏见以及一个新的方面：组织权力偏见。实验表明，DALLE-3生成的图像中超过74%表现出性别职业偏见。此外，与单人设置相比，DALLE-3在PST下更倾向于延续与男性相关的刻板印象。我们进一步提出了FairCritic，这是一个新颖且可解释的框架，利用基于大语言模型（LLM）的评判模型来i）检测生成图像中的偏见，以及ii）自适应地向T2I模型提供反馈以提升公平性。FairCritic在PST上实现了近乎完美的公平性，克服了以往基于提示干预方法的局限性。
英文摘要：
Recent large-scale T2I models like DALLE-3 have made progress in reducing gender stereotypes when generating single-person images. However, significant biases remain when generating images with more than one person. To systematically evaluate this, we propose the Paired Stereotype Test (PST) framework, which queries T2I models to depict two individuals assigned with male-stereotyped and female-stereotyped social identities, respectively (e.g. "a CEO" and "an Assistant"). This contrastive setting often triggers T2I models to generate gender-stereotyped images. Using PST, we evaluate two aspects of gender biases -- the well-known bias in gendered occupation and a novel aspect: bias in organizational power. Experiments show that over 74\% images generated by DALLE-3 display gender-occupational biases. Additionally, compared to single-person settings, DALLE-3 is more likely to perpetuate male-associated stereotypes under PST. We further propose FairCritic, a novel and interpretable framework that leverages an LLM-based critic model to i) detect bias in generated images, and ii) adaptively provide feedback to T2I models for improving fairness. FairCritic achieves near-perfect fairness on PST, overcoming the limitations of previous prompt-based intervention approaches.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
