1 paper
Zhuohong Chen, Zhengxian Wu, Zirui Liao +6
Vision-centric retrieval for VQA requires retrieving images to supply missing visual cues and integrating them into the reasoning process. However, selecting the right images and i…