1
SER: Learning to Ground Video Reasoning with Semantic Evidence Rewards
提出语义证据奖励机制，让视频推理模型学会“先找证据再回答”，精度与可解释性兼得。
arXiv:2606.24726v1 Announce Type: new Abstract: Video MLLMs often struggle with fine-grained spatio-temporal reasoning, sometimes generating correct a…