AI RESEARCH
Decoupling Skeleton and Flesh: Efficient Multimodal Table Reasoning with Disentangled Alignment and Structure-aware Guidance
arXiv CS.CL
•
ArXi:2602.03491v2 Announce Type: replace-cross Reasoning over table images remains challenging for Large Vision-Language Models (LVLMs) due to complex layouts and tightly coupled structure-content information. Existing solutions often depend on expensive supervised