AI Paper Atlas

Research map / Video / Video-language understanding

MMG-Vid: Maximizing Marginal Gains at Segment-level and Token-level for Efficient Video LLMs

Junpeng Ma, Qizhe Zhang, Ming Lu, Zhibin Wang, Qiang Zhou, Jun Song, Shanghang Zhang

AAAI 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 14 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 MMG-Vid pruning at 10% retention versus FastVLLaVA-Video-7B · Average over 4 benchmarks, 10% retentionrelative performance (%): 77.1 -> 93.8Improved
prune 75% of tokens with MMG-VidLLaVA-OneVision-7B · Average over 4 benchmarks, 25% retentionrelative performance (%): 100.0 -> 99.5Improved
use TG-DPC (temporal-guided) instead of standard DPC-KNN for token pruning (ablation)MMG-Vid · MVBench, 25% retentionrelative performance (%): 94.6 -> 97.3Improved
use marginal gain-based dynamic budget instead of uniform static budget (ablation)MMG-Vid · LongVideoBench and MLVU, 25% retentionrelative performance (%): +0.8% LongVideoBench, +1.0% MLVUImproved
use balanced lambda=0.5 instead of representativeness only (lambda=1.0) (ablation)MMG-Vid segment budgeting · Average, 25% retentionrelative performance (%): 96.6 -> 97.6Improved
use balanced lambda=0.5 instead of diversity only (lambda=0) (ablation)MMG-Vid segment budgeting · Average, 25% retentionrelative performance (%): 97.2 -> 97.6Improved
diversity-only budgeting versus representativeness-only budgeting (ablation)MMG-Vid segment budgeting · Average, 25% retentionrelative performance (%): 96.6 -> 97.2Improved
MMG-Vid versus FrameFusion at 25% retentionLLaVA-Video-7B · Average, 25% retentionrelative performance (%): 96.8 -> 98.1Improved
MMG-Vid versus VisionZip at 25% retention on LLaVA-OneVisionLLaVA-OneVision-7B · Average, 25% retentionrelative performance (%): 99.2 -> 99.5Improved

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