Skip to content
AI News HubLIVE
Original source2 min read

StreamScout: Learning When to Look Deeper for Streaming Video Understanding

Summary

StreamScout is an adaptive inference framework for streaming video understanding that keeps only a lightweight textual timeline and progressively escalates through three visual views per query, stopping as soon as the evidence suffices. Distillation and reinforcement learning refine this stop-or-escalate policy, improving accuracy while cutting tokens and latency.

SourcearXiv Computer VisionAuthor: Ce Zhang, Jing Bi, Jinxi He, Jianshu Zhang, Jingyang Lin, Yunzhong Xiao, Minghao Fu, Yaqi Xie, Zhentao Xie, Weicong Chen, Katia Sycara, Ming Zhou
StreamScout: Learning When to Look Deeper for Streaming Video Understanding
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 31 Aug 2026]

Title:StreamScout: Learning When to Look Deeper for Streaming Video Understanding

View a PDF of the paper titled StreamScout: Learning When to Look Deeper for Streaming Video Understanding, by Ce Zhang and 11 other authors

View PDF HTML (experimental)

Abstract:Streaming video understanding requires answering questions that arrive at arbitrary moments over an unbounded video stream. Existing systems primarily focus on what to retain in a bounded memory, yet access that memory using the same fixed-cost procedure for every query, despite substantial variation in the evidence required. We argue that deciding how deeply to access memory for each query is as important as deciding what the memory should store. To this end, we introduce StreamScout, an adaptive inference framework that maintains only a lightweight textual timeline in context as the stream unfolds. At query time, StreamScout progressively augments the timeline with up to three increasingly informative visual views: a glance at recent frames, a uniform look-back over the past stream, and query-salient retrieval. At each stage, the model answers immediately if the available evidence is sufficient; otherwise, it escalates to the next view. To improve this stop-or-escalate policy, we probe the cascade on an auxiliary set and distill the model's empirical competence boundary into supervision for a lightweight LoRA adaptation, yielding StreamScout-S. We further refine the policy through reinforcement learning, allowing the model to explore stopping behaviors beyond imitation of the distilled decisions, yielding StreamScout-R. Across three backbones and three streaming benchmarks, StreamScout and its variants consistently outperform prior streaming methods while substantially reducing inference cost and token consumption; on OVO-Bench, for instance, StreamScout-S improves Qwen3-VL-8B by 14.65 points while using 59% fewer tokens than uniform sampling and answering in 1.04 s on average.

Subjects:

Computer Vision and Pattern Recognition (cs.CV)

Cite as: arXiv:2609.00291 [cs.CV]

(or arXiv:2609.00291v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2609.00291

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Ce Zhang [view email] [v1] Mon, 31 Aug 2026 19:35:25 UTC (3,754 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled StreamScout: Learning When to Look Deeper for Streaming Video Understanding, by Ce Zhang and 11 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CV

new | recent | 2026-09

Change to browse by:

cs

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?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • StreamScout keeps a lightweight textual timeline and escalates through recent-frame glance, uniform look-back, and query-salient retrieval only when needed.
  • StreamScout-S distills the model's empirical competence boundary into LoRA supervision; StreamScout-R further refines the stopping policy with reinforcement learning.
  • On OVO-Bench, StreamScout-S improves Qwen3-VL-8B by 14.65 points with 59% fewer tokens than uniform sampling and 1.04 s average answering time.
  • Across three backbones and three streaming benchmarks, the method consistently outperforms prior streaming approaches at substantially lower inference cost.

Highlights and analysis are generated automatically and may contain errors. Check the original source.