AI Paper Atlas

Research map / Video / Video-language understanding

CogStream: Context-guided Streaming Video Question Answering

Zicheng Zhao, Kangyu Wang, Shijie Li, Rui Qian, Weiyao Lin, Huabin Liu

AAAI 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 4 citations (OpenAlex)

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
add temporal-semantic clustering (ablation)CogReasoner (fine-tuned) · CogStreamavg score: 71.48 -> 72.26Improved
summarize low-relevance events instead of discarding them (ablation)CogReasoner (fine-tuned) · CogStreamavg score: 70.52 -> 72.26Improved
add full visual stream compression vs baseline (ablation)CogReasoner (fine-tuned) · CogStreamavg score: 70.64 -> 72.26Improved
use all historical QA instead of none (ablation)CogReasoner (fine-tuned) · CogStreamavg score: 66.04 -> 70.48Improved
retrieve relevant historical QA instead of using all context (ablation)CogReasoner (fine-tuned) · CogStreamavg score: 70.48 -> 72.26Improved
use retrieved context instead of all context on Basic QA (ablation)CogReasoner (fine-tuned) · CogStream Basicscore: 75.20 -> 75.30No clear effect
use all history vs none on Basic QA (ablation)CogReasoner (fine-tuned) · CogStream Basicscore: 75.60 -> 75.20No clear effect
add trained HDR module (ablation)VideoLLaMA2 · CogStreamavg score: 50.72 -> 55.16Improved
add trained HDR module (ablation)MiniCPM-V-2.6 · CogStreamavg score: 66.84 -> 67.78Improved
let the model's own LLM do historical retrieval (ablation)VideoLLaMA2 · CogStreamavg score: 50.72 -> 50.52Hurt
let the model's own LLM do historical retrieval (ablation)MiniCPM-V-2.6 · CogStreamavg score: 66.84 -> 65.46Hurt
add 30% irrelevant QA noise with all-context (ablation)CogReasoner (fine-tuned) · CogStreamavg score: 70.48 -> 69.74Hurt

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