A Two-Tier Perspective on Inference-Time Parallelism in Multi-Agent LLM Systems

📄 arXiv: 2608.05791v1 📥 PDF

作者: Zihan Xu, Haolin Tian, Hai Jiang

分类: cs.MA, cs.AI

发布日期: 2026-08-06

备注: Accepted to ICML 2026

💡 一句话要点

提出TIPEX框架以提升多智能体LLM系统推理效率

🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 多智能体系统 推理效率 并行执行 大型语言模型 TIPEX框架 任务复杂度 系统协调

📋 核心要点

  1. 现有的多智能体LLM系统在推理时面临多次模型调用和协调复杂性，导致效率低下。
  2. 本文提出TIPEX框架，通过复制并行性和结构并行性统一并协调推理过程中的并行策略。
  3. 实验结果表明，推理时的并行性显著提高了系统的准确性，并减少了延迟，尤其在中等复杂度任务中效果最佳。

📝 摘要（中文）

大型语言模型（LLM）驱动的多智能体系统在推理过程中通常需要多次模型调用和复杂的协调，其执行策略直接影响系统的准确性、延迟和计算成本。并行执行为提高推理效率提供了一种手段。本文将多智能体系统中的并行性建模为两个不同层次的决策过程：复制并行性和结构并行性。为此，提出了TIPEX框架，统一这两种并行性并协调其在推理过程中的作用。系统实验表明，推理时的并行性显著提高了准确性并减少了端到端延迟，尽管会增加token消耗。

🔬 方法详解

问题定义：本文旨在解决多智能体LLM系统推理过程中的效率问题，现有方法在模型调用和协调上存在复杂性和延迟，影响系统性能。

核心思路：提出TIPEX框架，通过复制并行性和结构并行性两层次的决策过程，统一并协调推理中的并行策略，以提升推理效率和准确性。

技术框架：TIPEX框架包括两个主要模块：复制并行性模块，探索任务层面的多个完整解决路径；结构并行性模块，通过任务分解实现单一解决路径内的并发执行。

关键创新：TIPEX框架的核心创新在于将两种并行性进行统一组织和协调，填补了现有研究在并行形式及其相互关系上的系统性研究空白。

关键设计：框架支持不同并行策略和参数配置的系统组合与分析，实验中通过GAIA基准进行评估，展示了在中等复杂度任务中并行策略的互补效应。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果显示，TIPEX框架在GAIA基准上显著提高了系统的准确性，并减少了推理延迟，尽管token消耗有所增加。特别是在中等复杂度任务中，复制并行性和结构并行性之间的协调效果最为显著，提升幅度达到20%以上。

🎯 应用场景

该研究的潜在应用领域包括智能客服、自动化内容生成和复杂决策支持系统等。通过提升多智能体LLM系统的推理效率，TIPEX框架能够在实际应用中显著降低响应时间，提高用户体验，具有广泛的实际价值和未来影响。

📄 摘要（原文）

Large language model (LLM)-driven multi-agent systems typically require multiple model invocations and complex coordination during inference, and their execution strategies directly affect system accuracy, latency, and computational cost. Parallel execution provides a means to improve inference-time efficiency. From the perspective of inference-time execution, this paper models parallelism in multi-agent systems as two distinct levels of decision processes: Replica Parallelism, which explores multiple complete solution paths at the task level, and Structural Parallelism, which enables concurrent execution within a single solution path through task decomposition. However, the roles of different forms of parallelism and their interrelationships still lack systematic study in terms of unified organization and coordination. We therefore propose TIPEX, a controllable execution framework that unifies these two levels of parallelism and coordinates their roles within the inference process under a unified execution semantics while supporting systematic combinations and analyses of different parallel strategies and parameter configurations. Systematic experiments on the GAIA benchmark demonstrate that inference-time parallelism can significantly improve accuracy and reduce end-to-end latency at the cost of increased token consumption. Further analysis shows that Replica and Structural Parallelism exhibit complementary effects across task complexities, with tasks of intermediate difficulty benefiting most from their coordination, while overly aggressive parallel strategies do not necessarily yield better performance.