| title | StreamOPD: A Post-Training Recipe with Spatio-Temporal Cue Gating for Streaming Video Understanding |
|---|---|
| authors | Keming Wu, Baoyi Wang, Kaichen Zhang, Xiang An, Zuhao Yang, Sudong Wang, Haowei Zhu, Tingxuan Huang, Hongcheng Gao, Bin Wang |
| date | 2026-08-17 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2608.16320 |
| affiliation | UniX Lab |
| upvotes | 10 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | Streaming video understanding demands direct responses from the causally observed prefix of an unfolding video. Existing systems add inference-time memory, retrieval, and compression, yet a training-free sliding-window baseline already matches them. We therefore fix a memory-free recent-window protocol and ask how far post-training alone can go. Reinforcement learning with verifiable rewards fits this regime poorly, encouraging long ``think-then-answer'' generations, while on-policy distillation (OPD) supplies dense token-level teacher supervision on student trajectories but is stable only when both models train in thinking mode. These observations lead to StreamOPD, a recipe combining verifiable streaming-video data, thinking-mode OPD, and instruct-mode deployment. It raises StreamingBench from 77.9% to 83.9%---within 0.3 points of the 9B teacher---and improves OVO-Bench excluding its hallucination-detection subtask (HLD) by 9.1 points under unchanged inference. As a teacher-privilege extension, Spatio-Temporal CueGate (ST-CueGate) aggregates cue-versus-no-cue teacher likelihood ratios into a group-relative response score that reweights OPD. It reaches 71.9% on OVO-Bench (excluding HLD) and 64.9% on Video-MME, and is the only variant that stays above the base model on all four benchmarks. Replacing the teacher with a frozen copy of the student's initial policy---on-policy self-distillation---retains most of these gains and lifts HLD to 57.0%, above both the untrained student and the 9B teacher, so abstention loss is not intrinsic to the recipe. We provide a transparent and reproducible reference for open-source streaming-video research. |
cite
@misc{wu2026streamopd,
title={StreamOPD: A Post-Training Recipe with Spatio-Temporal Cue Gating for Streaming Video Understanding},
author={Keming Wu and Baoyi Wang and Kaichen Zhang and Xiang An and Zuhao Yang and Sudong Wang and Haowei Zhu and Tingxuan Huang and Hongcheng Gao and Bin Wang},
year={2026},
eprint={2608.16320},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2608.16320}
}