---
id: 20261028-T0-15
title: "全模态大模型推理提速：按查询内容只调用相关模态，降低计算开销"
title_en: "Per-Query Modality Selection Cuts Omni-Modal LLM Inference Costs"
url: https://ai.daily.yangsir.net/daily/20261028-T0-15
issue_date: 2026-10-28
publish_date: 2026-10-27T04:00:00.000Z
category: research
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2610.31768
---

# 全模态大模型推理提速：按查询内容只调用相关模态，降低计算开销

arXiv新论文提出一种全模态大模型推理优化方案，核心思路是按查询内容动态选择所需模态，而非每次都处理文本、音频、视频的全部输入。研究观察到，全模态模型虽能联合处理多种模态，但推理成本高昂，而多数查询实际只涉及其中一两种。该方法通过per-query模态选择机制跳过无关模态的计算，在保持理解与推理能力的前提下降低开销。论文编号arXiv:2610.31768v1，属于cs.LG方向。实际影响：多模态应用部署方可以在不牺牲响应质量的情况下压缩推理成本，尤其适合视频、音频混合输入的实时场景。

## English Version

**Per-Query Modality Selection Cuts Omni-Modal LLM Inference Costs**

A new arXiv paper proposes an inference optimization for omni-modal large language models that dynamically selects which modalities a query actually needs, instead of processing text, audio, and video inputs every time. The authors observe that while omni-modal models can jointly handle many modalities, inference costs are high and most queries touch only one or two of them. The per-query modality selection mechanism skips computation on irrelevant modalities while preserving understanding and reasoning quality. The paper is arXiv:2610.31768v1 in cs.LG. In practice, teams deploying multimodal applications can cut inference costs without sacrificing response quality, especially in real-time scenarios with mixed video and audio input.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2610.31768)

**详情页**：https://ai.daily.yangsir.net/daily/20261028-T0-15

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*