AI Paper Atlas

Research map / Video / Video-language understanding

QueryStream: Advancing Streaming Video Understanding with Query-Aware Pruning and Proactive Response

Kairui Zhang, Zhenyu Yang, Bing Wang, Shengsheng Qian, Changsheng Xu

ICLR 2026 · Video › Video-language understanding · approach: Long video understanding

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace query-agnostic DTD with query-aware QDP at ~55% keep rateTimeChat-Online-7B · StreamingBenchoverall accuracy: 74.32 -> 75.32Improved
replace DTD with QDP at ~30% keep rateTimeChat-Online-7B · StreamingBenchoverall accuracy: 72.96 -> 74.04Improved
replace DTD with QDPTimeChat-Online-7B · OVO-Benchoverall accuracy: 46.7 (full) -> 49.4Improved
QueryStream at 20.0% keep vs TimeChat-Online at 55.4% keepTimeChat-Online-7B · OVO-Benchoverall accuracy: 47.6 -> 47.5Mixed
apply QDP zero-shot to Qwen2.5-VL (52.4% keep)Qwen2.5-VL-7B · VideoMMEoverall accuracy: 63.2 -> 63.6Improved
apply QDP to Qwen2.5-VL on long videosQwen2.5-VL-7B · VideoMME longaccuracy: 50.4 -> 52.6Improved
prune to 16.6% keep rate with QueryStreamTimeChat-Online-7B · LongVideoBenchaccuracy: 57.3 -> 58.0Improved
add query-aware QDP at ~57% keep rate versus no pruningTimeChat-Online-7B · StreamingBenchoverall accuracy: 75.36 -> 75.32No clear effect
combine semantic and temporal filters rather than temporal filter alone (ablation)QDP · StreamingBenchoverall accuracy: 74.76 -> 75.32Improved
combine semantic and temporal filters rather than semantic filter alone (ablation)QDP · StreamingBenchoverall accuracy: 74.52 -> 75.32Improved
add density gate to relevance trigger (full RTAR) (ablation)RTAR · OVO-Bench Forward Active Respondingtimeliness score: 30.2 -> 34.6Improved
add relevance gate to density-only trigger (ablation)RTAR · OVO-Bench Forward Active Respondingavg accuracy: 36.8 -> 40.2Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy