StreamPro: From Reactive Perception to Proactive Decision-Making in Streaming Video
Proactive streaming video understanding requires models to decide when to respond under partial observations, not just what to respond. Existing benchmarks follow a 'see-then-answer' paradigm, failing to evaluate timely and reliable decision-making with incomplete information. This paper introduces StreamPro-Bench, which evaluates models from three perspectives: Perception Understanding, Temporal Reasoning, and Proactive Agency. It also proposes StreamPro, a two-stage training framework using CB-Stream Loss and GRPO with multi-grained rewards, achieving 41.5 on StreamPro-Bench (vs. previous best 10.4) and 78.9 on StreamingBench-RTVU.
[2605.16381] StreamPro: From Reactive Perception to Proactive Decision-Making in Streaming Video
[Submitted on 11 May 2026]
Title:StreamPro: From Reactive Perception to Proactive Decision-Making in Streaming Video
View a PDF of the paper titled StreamPro: From Reactive Perception to Proactive Decision-Making in Streaming Video, by Ao Li and 9 other authors
View PDF HTML (experimental)
Abstract:Proactive streaming video understanding requires models to continuously process video streams and decide when to respond, rather than merely what to respond. This naturally introduces a decision-making problem under partial observations, where models must balance early prediction against sufficient evidence. However, existing benchmarks largely follow a "see-then-answer" paradigm, where responses are triggered only after explicit evidence appears, effectively reducing proactive reasoning to delayed perception. As a result, they fail to evaluate a model's ability to make timely and reliable decisions under incomplete observations. Moreover, training proactive models is inherently challenging due to the extreme imbalance between silence and response signals in streaming trajectories, as well as the need to jointly optimize response correctness and timing. To address these challenges, we introduce StreamPro-Bench, a new benchmark that evaluates streaming models from three complementary perspectives: Perception Understanding, Temporal Reasoning, and Proactive Agency, where the last measures a model's ability to make early yet reliable decisions under partial observations. We further propose StreamPro, a two-stage training framework for proactive learning. First, we introduce CB-Stream Loss to mitigate the severe supervision imbalance during supervised fine-tuning (SFT). Then, we apply Group Relative Policy Optimization (GRPO) with a multi-grained reward design that involves both turn-level and trajectory-level rewards. Experiments show that StreamPro significantly improves proactive performance. On StreamPro-Bench, it achieves 41.5, substantially outperforming the previous best (10.4), while also maintaining strong performance on real-time streaming benchmarks, achieving 78.9 on StreamingBench-RTVU.
Subjects:
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Cite as: arXiv:2605.16381 [cs.CV]
(or arXiv:2605.16381v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2605.16381
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Ao Li [view email] [v1] Mon, 11 May 2026 05:01:15 UTC (4,651 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled StreamPro: From Reactive Perception to Proactive Decision-Making in Streaming Video, by Ao Li and 9 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.CV
new | recent | 2026-05
Change to browse by:
cs cs.AI
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
Loading...
Data provided by:
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)