一作ACL 2026Poster· Natural Language Processing
Measuring Social Bias in Vision-Language Models with Face-Only Counterfactuals from Real Photos
Haodong Chen · Qiang Huang · Jiaqi Zhao · Qiuping Jiang · Xiaojun Chang · Jun Yu
摘要
Vision-Language Models (VLMs) are increasingly deployed in socially consequential settings, raising concerns about social bias driven by demographic cues. A central challenge in measuring such social bias is attribution under visual confounding: real-world images entangle race and gender with cor...