Chaoya Jiang

IDAICL improves predictions by refining demonstration quality in in-context learning.

MIBench tests multimodal models' performance on multiple images.

MaVEn enhances AI's ability to process multiple images for better reasoning.