PulseAugur
实时 11:13:32
English(EN) Beyond One-Size-Fits-All: Sample-Adaptive Strategy Routing for Vision Token Pruning in MLLMs

新的VIP-Router通过自适应策略选择优化MLLM视觉Token剪枝

研究人员开发了VIP-Router,一个新颖的系统,旨在通过为每个输入自适应地选择最佳视觉Token剪枝策略来优化多模态大语言模型(MLLM)的效率。与之前在所有输入上应用单一策略的方法不同,VIP-Router分析低成本的视觉和文本特征,以预测哪种剪枝方法将产生最高的准确性和效用。这个即插即用(plug-and-play)的解决方案与现有的MLLM和剪枝算法无缝集成,引入了最少的训练参数。在VTC-Bench Group A基准上的评估表明,VIP-Router显著优于固定策略基线,在平均准确率上实现了26.9%的相对提升,在平均效用上实现了22.0%的相对增长。 AI

影响 通过自适应地选择最佳视觉Token剪枝策略来提高MLLM的效率,有可能降低推理成本并改善各种模型的性能。

排序理由 详细介绍优化MLLM新方法的论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CV 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

新的VIP-Router通过自适应策略选择优化MLLM视觉Token剪枝

本文如何被排名

Signal score
10 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
详细介绍优化MLLM新方法的论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
paper, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Same-day
Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准

报道来源 [1]

  1. arXiv cs.CV TIER_1 English(EN) · Haiji Liang, Pengfei Zhou, Zhenglin Wan, Wei Wang, Yang You, Wangbo Zhao ·

    超越“一刀切”:用于多模态大模型视觉令牌剪枝的样本自适应策略路由

    arXiv:2609.10346v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) process hundreds or thousands of visual tokens per image, incurring prohibitive inference costs. While existing vision token pruning methods mitigate this overhead, they implicitly assume tha…