#multimodal-reasoning

CURV targets visual grounding in chart question answering

The arXiv paper introduces a curriculum framework and CCQA dataset for multimodal chart reasoning.