AI Paper Atlas

Research map / Video / Video-language understanding

Towards Effective Long Video Understanding: Dynamic MAS Construction via Meta-Agent

Jing Huang, Lidong Zhang, Yadong Li, Xingzhong Xu, Siye Chen, Jie Liu, Ming Kong, Qiang Zhu

ECCV 2026 · Video › Video-language understanding · approach: Long video understanding · 0 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 SFT on synthesized MAS code (ablation)DyMAC MAS-Visual · MLVU/Video-MME/LVB averageaccuracy: 51.2 -> 65.5Improved
add RLVR after SFT (ablation)DyMAC MAS-Visual · MLVU/Video-MME/LVB averageaccuracy: 65.5 -> 68.2Improved
give Meta-Agent the video as well as the query (ablation)DyMAC after RL · MLVU/Video-MME/LVB averageaccuracy: 64.4 -> 68.2Improved
add SFT to text-only Meta-Agent (ablation)DyMAC MAS-Text · MLVU/Video-MME/LVB averageaccuracy: 50.9 -> 63.1Improved
untrained dynamic MAS vs vanilla baseline (ablation)DyMAC MAS-Visual · MLVU/Video-MME/LVB averageaccuracy: 63.3 -> 51.2Hurt
add accuracy reward in RLVR (ablation)DyMAC · MLVUaccuracy: 70.9 -> 72.7Improved
add efficiency reward to accuracy reward (ablation)DyMAC · Inference timeseconds: 13.2 -> 4.6Improved
add efficiency reward to accuracy reward (ablation)DyMAC · MLVUaccuracy: 72.7 -> 73.2No clear effect
dynamic 3b/7b driving model allocation instead of 3b only (ablation)DyMAC · MLVUaccuracy: 69.7 -> 73.2Improved
dynamic 3b/7b allocation instead of 7b only (ablation)DyMAC · Inference timeseconds: 9.8 -> 4.6Improved
add CoT to reasoning patterns (ablation)DyMAC · MLVU/Video-MME/LVB averageaccuracy: 64.5 -> 65.7Improved
add LLM-Debate to reasoning patterns (ablation)DyMAC · MLVU/Video-MME/LVB averageaccuracy: 65.7 -> 66.6Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 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