跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2604.03072 · v2

MI-Pruner: Crossmodal Mutual Information-guided Token Pruner for Efficient MLLMs

首次提交 2026-04-03 22:51
版本更新 2026-11-16 16:03
雷达观测 2026-11-17 17:11
arXiv ↗PDF ↗

摘要

The redundancy of visual tokens in multimodal large language models (MLLMs) motivates training-free token pruning techniques for efficient inference. Existing methods typically rank visual tokens by attention salience and retain only the most salient ones. However, salience-based pruning is susceptible to sink tokens and positional bias, overlooking the diversity of the selected set. Moreover, extracting attention maps incurs additional overhead. In this paper, we propose MI-Pruner, which derives Pointwise Mutual Information (PMI) based relevance scores in the projection space to reflect crossmodal relevance and intra-modal redundancy. Grounded in information theory, our approach identifies the most informative tokens by explicitly measuring feature-level dependency. Without reliance on internal attention maps or architectural modifications, MI-Pruner can be seamlessly applied to any pretrained MLLMs for inference acceleration. Extensive experiments on LLaVA1.5 and Qwen-series demonstrate that our approach achieves a favorable performance-efficiency trade-off. Specifically, we preserve >97%{>}97\% performance on LLaVA1.5-7B at a pruning rate of 88.9\%.

作者备注:9 pages

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-16 16:03 · 最新观测版本

MI-Pruner: Crossmodal Mutual Information-guided Token Pruner for Efficient MLLMs

The redundancy of visual tokens in multimodal large language models (MLLMs) motivates training-free token pruning techniques for efficient inference. Existing methods typically rank visual tokens by attention salience and retain only the most salient ones. However, salience-based pruning is susceptible to sink tokens and positional bias, overlooking the diversity of the selected set. Moreover, extracting attention maps incurs additional overhead. In this paper, we propose MI-Pruner, which derives Pointwise Mutual Information (PMI) based relevance scores in the projection space to reflect crossmodal relevance and intra-modal redundancy. Grounded in information theory, our approach identifies the most informative tokens by explicitly measuring feature-level dependency. Without reliance on internal attention maps or architectural modifications, MI-Pruner can be seamlessly applied to any pretrained MLLMs for inference acceleration. Extensive experiments on LLaVA1.5 and Qwen-series demonstrate that our approach achieves a favorable performance-efficiency trade-off. Specifically, we preserve >97%{>}97\% performance on LLaVA1.5-7B at a pruning rate of 88.9\%.

查看此版本解读 →查看此版本原文 ↗