AI Paper Atlas

Research map / Video / Video-language understanding

MARC: Memory-Augmented RL Token Compression for Efficient Video Understanding

Peiran Wu, Zhuorui Yu, Yunze Liu, Chi-Hao Wu, Enmin Zhou, Junxiao Shen

ICLR 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 5 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 Visual Memory Retriever (top-3 clips) to Qwen2.5-VL-3B 64 frames (ablation)Qwen2.5-VL-3B · 6 video benchmarks (mean)accuracy: 42.21 -> 45.56Improved
add VMR to Qwen2.5-VL-3B on VideoMME (ablation)Qwen2.5-VL-3B · VideoMME (w/o sub)accuracy: 53.55 -> 51.85Hurt
add VMR to Qwen2.5-VL-3B on VideoMMMU (ablation)Qwen2.5-VL-3B · VideoMMMUaccuracy: 35.33 -> 34.33Hurt
use C-GRPO instead of SFT for 1-frame compression (ablation)MARC-3B · 6 video benchmarks (mean)accuracy: 38.50 -> 42.20Improved
use C-GRPO instead of SFT, both with VMR (ablation)MARC-3B · 6 video benchmarks (mean)accuracy: 40.16 -> 42.20Improved
add VMR to SFT-trained 1-frame model (ablation)Qwen2.5-VL-3B SFT · 6 video benchmarks (mean)accuracy: 38.50 -> 40.16Improved
set retention threshold tau 0.6 instead of 0.4 (ablation)MARC-3B · 6 video benchmarks (mean)accuracy: 41.34 -> 42.20Improved
set retention threshold tau 0.6 instead of 0.8 (ablation)MARC-3B · 6 video benchmarks (mean)accuracy: 41.40 -> 42.20Improved
compress to 1 frame-equivalent tokens with MARC vs 64-frame baselineMARC-3B · VideoMME (w/o sub)accuracy: 53.55 -> 39.44Hurt
compress to 1 frame-equivalent tokens with MARC vs 64-frame baselineMARC-3B · VSI-Benchaccuracy: 32.93 -> 27.55Hurt
compress to 1 frame-equivalent tokens with MARC vs 64-frame baselineMARC-3B · TempCompassaccuracy: 38.05 -> 55.34Improved
compress to 1 frame-equivalent tokens with MARC vs 64-frame baselineMARC-3B · 6 video benchmarks (mean)accuracy: 42.21 -> 42.20No clear effect

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