作者位次 #3/5ACL 2026Poster· Natural Language Processing
V ideo C u RL : Video Curriculum Reinforcement Learning with Orthogonal Difficulty Decomposition
Hongbo Jin · Kuanwei Lin · Wenhao Zhang · Yichen Jin · Ge Li
摘要
Reinforcement Learning (RL) is crucial for empowering Video-LLMs with complex spatiotemporal reasoning. However, current RL paradigms predominantly rely on random data shuffling or naive curriculum strategies based on scalar difficulty metrics. We argue that scalar metrics fail to disentangle two...