SANER: Annotation-free Societal Attribute Neutralizer for Debiasing CLIP

Documenti analoghi

• From Descriptive Richness to Bias: Unveiling the Dark Side of Generative Image Caption Enrichment
  di: Hirota, Yusuke, et al.
  Pubblicazione: (2024)
• Interpretable Debiasing of Vision-Language Models for Social Fairness
  di: An, Na Min, et al.
  Pubblicazione: (2026)
• From Global to Local: Social Bias Transfer in CLIP
  di: Ramos, Ryan, et al.
  Pubblicazione: (2025)
• Resampled Datasets Are Not Enough: Mitigating Societal Bias Beyond Single Attributes
  di: Hirota, Yusuke, et al.
  Pubblicazione: (2024)
• LOTUS: A Leaderboard for Detailed Image Captioning from Quality to Societal Bias and User Preferences
  di: Hirota, Yusuke, et al.
  Pubblicazione: (2025)