arXiv cs.CL· Xueyun Tian, Bingbing Xu, Jiaming Liu, Jinpeng Yu, Yang Shi, Yu Zhao, Ruihua Huang, Yuanzhuo Wang, Huawei Shen·· 12 小时前AI 评分32
SyncPro：面向主动式流视频理解的同步推理与响应框架
SyncPro: Synchronous Reasoning and Responding for Proactive Streaming Video Understanding
AI 导读
SyncPro 是一个同步双流框架，用 Inner 与 Response 两条独立生成路径分别负责证据追踪和面向用户的响应，两者共享因果历史、可独立激活，使响应能与持续推理并行进行。研究团队构建了 189K 双流样本，经监督微调加轨迹级强化学习训练，在三个主动式基准上提升了响应时机与回答准确率。
正文 · 原文
Abstract:Proactive streaming video understanding requires assistants to track unfolding events and respond when a user's requested condition is met. To make these decisions, assistants must accumulate relevant evidence and reassess response conditions as new observations arrive. Existing work offers complementary capabilities for this process: proactive video assistants learn when to respond, while streaming video reasoning methods generate scheduled intermediate text to organize observations and support subsequent answers. Together, these capabilities provide a foundation for proactive reasoning, but how to coordinate ongoing reasoning with timely responses remains underexplored. In a direct adaptation, fixed reasoning intervals can trigger updates when nothing relevant has changed or postpone useful updates until the next scheduled step. Sharing one generation path also couples two distinct requirements: preserving evidence for future decisions and producing only the response requested by the user. Moreover, responses generated after reasoning must wait for it to finish, even when sufficient evidence is already available. To address these challenges, we introduce SyncPro, a synchronous dual-stream framework with specialized generation paths for evidence tracking and user-facing responses. Its Inner and Response branches share causal history and activate independently, allowing relevant evidence changes to trigger internal updates and responses to proceed concurrently with ongoing reasoning. We construct 189K dual-stream examples and train SyncPro through supervised fine-tuning followed by trajectory-level reinforcement learning. Experiments across three proactive benchmarks show improved response timing and more accurate answers to user requests.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.32132 [cs.CV] |
| (or arXiv:2610.32132v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.32132 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Xueyun Tian [view email]
[v1]
Fri, 23 Oct 2026 03:58:17 UTC (20,505 KB)
来源：arXiv cs.CL · arxiv.org