AI RESEARCH

Reveal-to-Revise: Explainable Bias-Aware Generative Modeling with Multimodal Attention

arXiv CS.AI

ArXi:2510.12957v2 Announce Type: replace-cross We present an explainable, bias-aware generative framework that unifies cross-modal attention fusion, Grad-CAM++ attribution, and a Reveal-to-Revise feedback loop within a single