Video-OPSD introduces a post‑training framework for Video Large Language Models that leverages privileged visual evidence to enhance on‑policy self‑distillation. The method constructs a self‑teacher conditioned only on annotated evidence frames, while the student processes the full video, allowing the teacher to provide more focused supervision. Additionally, an evidence‑guided token optimization scheme weights distillation based on each token’s reliance on privileged evidence, improving perceptually grounded reasoning. Experiments demonstrate consistent gains over standard OPSD and comparable performance to GRPO with less training time.
By Ziyue Wang, Shiqi Huang, Weiwen Xu, Bihan Wen, Xudong Jiang
arXiv:2605. 06094v5 Announce Type: replace-cross Abstract: Training VideoLLMs for complex reasoning remains challenging due to sparse sequence level rewards and the lack of fine grained credit assignment over long, temporally grounded reasoning trajectories.
By Hao Lin, Kunyang Lv, Xu Jiang, Jingqi Tian, Zhongjing Du, Jiayu Ding, Qiaoman Zhang, Hongbo Jin
arXiv:2608. 16316v1 Announce Type: cross Abstract: Large Multimodal Models (LMMs) for video reasoning have long been hindered by the high computational cost of processing vast amounts of visual information.
By Ao Shen, Yongheng Zhang, Yinghui Li, Manning Wang, Di Yin, Xing Sun
The paper introduces OWLET, an on-policy weighted learning method for post-training Video Large Language Models. OWLET leverages a richer teacher view that combines global temporal context with question-conditioned local windows to provide visual evidence, and it employs evidence-aware credit redistribution and spatiotemporal state alignment to better utilize this evidence during reinforcement learning. The approach aims to improve video reasoning by aligning token-level supervision with spatiotemporal representations.
By Qitan Lv, Ke Fan, Hong Wang, Wen Wu, Yilin Wang, Bowen Zhou, Feng Wu, Chao Zhang
arXiv:2609.40055v1 Announce Type: cross
Abstract: On-policy distillation (OPD) provides dense supervision directly on student-generated trajectories, making it an effective post-training strategy for...
By Jiacheng Qiu, Yunsoo Kim, Ruichen Xu, Jian Luo, Petar M. Djuri\'c, Sima Mofakham
arXiv:2610.14939v1 Announce Type: new
Abstract: GRPO-style reinforcement learning has advanced vision-language reasoning, but its sequence-level rewards provide limited credit assignment for individu...
By Jiahe Feng
DyMD introduces a Distribution Matching Distillation framework that adapts teacher supervision and critic fitting to preserve interaction dynamics in few-step video generation. By employing temporal affinity–conditioned re‑noise sampling and dynamics‑guided fake‑score tracking, DyMD balances motion recovery with visual quality. The method distills a 14B teacher into a 1.3B student that achieves significant gains on embodied‑video benchmarks and downstream action planning tasks.
By Haojun Xu, Jie Huang, Xin Lu, Mingchen Zhong, Zihao Fan, Linjiang Huang, Si Liu
The paper introduces LARVA, a framework that distills multi‑round video agents by jointly learning latent reasoning and visual observations. Instead of using textual captions, an observer sends compact continuous representations to a reasoner, which then selects the next action or produces an answer. The approach incorporates action and answer supervision through this interface and uses observation reconstruction and reasoning‑distribution alignment as auxiliary signals, achieving strong results on five video‑understanding benchmarks with 4B and 7B models.
By Junzhang Liu, Hao Yuan, Qi Zhang, Yiming Wu, Long Chen, Zhecan Wang
arXiv:2609.34581v2 Announce Type: replace
Abstract: Temporal video grounding is a key capability of advanced Multimodal Large Language Models (MLLMs) for the thorough understanding of video events, w...
By Shaobo Ju, Haiyang Yu, Xuecheng Wu, Qiong Wu, Jiacong Wang, Fan Shi, Jun Peng, Yiyi Zhou
The paper introduces Clue-OPSD, a clue‑privileged on‑policy self‑distillation framework that improves long‑video understanding by focusing on short, question‑relevant clue intervals rather than the entire video. Experiments on multiple benchmarks and Qwen3.5 model scales show that this approach consistently outperforms standard backbone models and competes strongly with supervised post‑training baselines, all while requiring fewer input frames and no additional inference modules.
By Kaishen Wang, Dongdi Zhao, Yijun Liang, Dingqiang Ye, Ruibo Chen, Heng Huang, Di Fu
arXiv:2610.16172v1 Announce Type: cross
Abstract: Recent multimodal large language models (MLLMs) achieve strong results on video question answering, yet often fail to recover the temporal structure...
By Guowen Zhang, Boshen Xu, Zihao Yue, Ziheng Wang, Xiaokun Liu, Xin Tao, Wenyu Qin, Pengfei Wan, Qin Jin
arXiv:2608.22364v1 Announce Type: new
Abstract: World action models (WAMs) couple visual future prediction with robot action generation, but accelerated students can lose task capabilities during dis...
By Liuhaichen Yang, Zhuang Jiang, Chenchao Sheng, Zezhi Tang