1 paper
Xianda Zheng, Huan Gao, Meng-Fen Chiang +3
Despite recent progress, the reasoning capabilities of large multimodal language models (MLLMs) remain fundamentally constrained by static supervision, where fixed prompts, rules,…