VisBias：衡量视觉语言模型中的显性和隐性社会偏见
VisBias: Measuring Explicit and Implicit Social Biases in Vision Language Models
AI总结：
本研究提出VisBias基准，通过直接问答和辅助任务分别测量视觉语言模型中的显性与隐性社会偏见，并评估多个主流模型，公开代码与数据。
摘要：
本研究调查了视觉语言模型（VLM）所表现出的显性和隐性社会偏见。这两种偏见类型的关键区别在于意识水平：显性偏见指有意识的、故意的偏见，而隐性偏见则在潜意识层面运作。为了分析显性偏见，我们直接向VLM提出与性别和种族差异相关的问题：（1）基于给定图像的多项选择题（例如，“图像中人物的教育水平是什么？”）（2）使用两张图像的“是-否”比较（例如，“第一张图像中的人物是否比第二张图像中的人物受教育程度更高？”）。对于隐性偏见，我们设计了VLM辅助用户但通过其回答揭示偏见的任务：（1）图像描述任务：要求模型描述图像中的人物，我们分析不同人口统计群体之间文本线索的差异。（2）表单填写任务：模型起草一份包含20个属性的个人信息收集表，我们检查所选属性之间的相关性以寻找潜在偏见。我们评估了Gemini-1.5、GPT-4V、GPT-4o、LLaMA-3.2-Vision和LLaVA-v1.6。我们的代码和数据可在https://github.com/uscnlp-lime/VisBias公开获取。
英文摘要：
This research investigates both explicit and implicit social biases exhibited by Vision-Language Models (VLMs). The key distinction between these bias types lies in the level of awareness: explicit bias refers to conscious, intentional biases, while implicit bias operates subconsciously. To analyze explicit bias, we directly pose questions to VLMs related to gender and racial differences: (1) Multiple-choice questions based on a given image (e.g., "What is the education level of the person in the image?") (2) Yes-No comparisons using two images (e.g., "Is the person in the first image more educated than the person in the second image?") For implicit bias, we design tasks where VLMs assist users but reveal biases through their responses: (1) Image description tasks: Models are asked to describe individuals in images, and we analyze disparities in textual cues across demographic groups. (2) Form completion tasks: Models draft a personal information collection form with 20 attributes, and we examine correlations among selected attributes for potential biases. We evaluate Gemini-1.5, GPT-4V, GPT-4o, LLaMA-3.2-Vision and LLaVA-v1.6. Our code and data are publicly available at https://github.com/uscnlp-lime/VisBias.
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。
