AI Paper Atlas

Research map / Video / Video-language understanding

EVA: Efficient Reinforcement Learning for End-to-End Video Agent

Yaolun Zhang, Ruohui Wang, Jiahao Wang, Yepeng Tang, Xuanyu Zheng, Haonan Duan, Hao Lu, Hanming Deng, Lewei Lu

CVPR 2026 · Video › Video-language understanding · approach: Long video 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 KTO stage after SFT (ablation)EVA · LongVideoBenchaccuracy (%): 49.9 -> 53.2Improved
add KTO stage after SFT (ablation)EVA · MLVUaccuracy (%): 52.3 -> 57.4Improved
add KTO stage after SFT (ablation)EVA · LVBenchaccuracy (%): 26.5 -> 36.0Improved
add KTO stage after SFT (ablation)EVA · VideoMME-Overallaccuracy (%): 56.0 -> 56.5No clear effect
add KTO stage after SFT (ablation)EVA · VideoMME-Longaccuracy (%): 45.8 -> 45.1Hurt
add GRPO stage after KTO (ablation)EVA · LongVideoBenchaccuracy (%): 53.2 -> 55.0Improved
add GRPO stage after KTO (ablation)EVA · MLVUaccuracy (%): 57.4 -> 68.3Improved
add GRPO stage after KTO (ablation)EVA · VideoMME-Overallaccuracy (%): 56.5 -> 60.2Improved
add GRPO stage after KTO (ablation)EVA · LVBenchaccuracy (%): 36.0 -> 43.3Improved
add GRPO stage after KTO (ablation)EVA · Video-Holmes (zero-shot)overall accuracy (%): 32.9 -> 37.2Improved
mix open-ended QA with multiple-choice QA in GRPO data (ablation)EVA-GRPO · VideoMMEaccuracyImproved
use EVA agent instead of uniform 32-frame samplingQwen2.5-VL-7B · LSDBenchaccuracy (%): 49.2 -> 51.0Improved

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