English(EN) MiCo: Mutual Information Coverage Optimization through Semantic Erasure Modeling for Efficient MLLM Inference

MiCo方法通过减少视觉标记来优化MLLM推理

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-01 04:00

研究人员开发了MiCo，一种新颖的无需训练的方法，通过减少处理的视觉标记数量来优化多模态大语言模型（MLLM）的推理。MiCo采用两阶段剪枝方法，基于语义擦除建模和任务感知子集选择来选择代表性的视觉标记。该方法在各种MLLM和基准测试中始终优于现有技术，在保持高性能的同时显著提高了推理速度。例如，MiCo将LLaVA-NEXT-13B的视觉标记减少到5.6%，实现了3.8倍的加速，而性能仅下降2.5%。 AI

影响该方法可以显著降低多模态人工智能系统的计算成本并提高其效率。

排序理由该集群描述了arXiv论文中提出的一种用于优化多模态大语言模型推理的新方法。[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 →

本文如何被排名

Signal score: 22 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群描述了arXiv论文中提出的一种用于优化多模态大语言模型推理的新方法。[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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CV TIER_1 English(EN) · Tinghao Wang, Yichen Guo, Qizhe Zhang, Yuan Zhang, Weimin Ouyang, Rui Huang, Jiajun Cao, Sixiang Chen, Hao Jiang, Jixian Wu, Zheng Lu, Bofan Zhu, Renyuan Li, Shanghang Zhang · 2026-10-01 04:00

MiCo：通过语义擦除建模实现互信息覆盖优化，以实现高效 MLLM 推理

arXiv:2609.34330v2 Announce Type: replace Abstract: Multimodal large language models (MLLMs) have demonstrated impressive performance in multimodal understanding, but processing large numbers of visual tokens results in high computational costs. While many methods have been propo…

本文如何被排名

报道来源 [1]

MiCo：通过语义擦除建模实现互信息覆盖优化，以实现高效 MLLM 推理

相关实体

相关话题