AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

TOP-RL: Task-Optimized Progressive Token Pruning with Reinforcement Learning for Vision Language Models

Hengyi Wang, Weiying Xie, Hui Jiang, Yaotao Wei, Kai Jiang, Mingxiang Cao, Chenhe Hao, Leyuan Fang

AAAI 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Attention-based visual token pruning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes TOP-RL, a task-optimized progressive visual token pruning method for vision-language models that uses reinforcement learning to decide pruning across tasks and layer depths.

Problem

LVLM inference is expensive because of the large number of visual tokens; existing compression methods prune at the input or in early layers and ignore variation across tasks and depths.

Approach

Progressive, task-aware visual token pruning across model layers, with the pruning policy optimized by reinforcement learning rather than fixed input-stage or early-layer rules.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use TOP-RL task-optimized progressive pruning instead of FastVLLaVA-1.5-7B, 192 tokens · 9-benchmark averageavg accuracy retention: 87.7% -> 97.6%Improved
use TOP-RL instead of VisionZip at 64 tokensLLaVA-1.5-7B · 9-benchmark averageavg accuracy retention: 87.9% -> 91.0%Improved
use TOP-RL on Video-LLaVA instead of VisionZipVideo-LLaVA · Video QA averageavg accuracy retention: 93.4% -> 93.9%Improved
use TOP-RL instead of VisionZip for efficiencyLLaVA-NeXT-7B · Efficiency (320 tokens)speedup: 2.11x -> 1.91xHurt
use TOP-RL instead of VisionZip for accuracy at same budgetLLaVA-NeXT-7B · Efficiency (320 tokens)F1 score: 86.4 -> 87.2Improved
transfer pruning agent across cognitive levels (perception to reasoning) (ablation)TOP-RL agent · SQAaccuracy: 69.6 -> 46.4Hurt
transfer pruning agent within the same cognitive domain (POPE agent to TextVQA) (ablation)TOP-RL agent · TextVQAaccuracy: 58.2 -> 58.9No clear effect
apply progressive multi-stage pruning (FastV-P) instead of one-shot FastV (ablation)FastV · TextVQA, SQAaccuracyImproved

Earlier methods this paper compares with

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