递归多智能体系统

摘要

近日，递归或循环语言模型通过潜在状态上的迭代计算深化推理，成为模型规模扩展的新方向。我们将这种扩展原理从单一模型推广至多智能体系统，并提出核心问题：智能体协作能否通过递归实现规模扩展？为此，我们提出递归多智能体框架RecursiveMAS，将整个系统构建为统一潜在空间内的递归计算。该框架通过轻量级递归链接模块将异构智能体连接为协作循环，实现分布内潜在思维的生成与跨智能体潜在状态传递。为优化框架性能，我们开发内外双循环学习算法，通过递归轮次间基于梯度的共享信用分配，实现全系统迭代协同优化。对运行时复杂度与学习动态的理论分析表明，RecursiveMAS较传统基于文本的多智能体系统更具效率，并能保持递归训练中的梯度稳定性。实验方面，我们在4种典型智能体协作模式下实例化RecursiveMAS，并在涵盖数学、科学、医学、搜索和代码生成的9个基准测试中评估。与先进单/多智能体及递归计算基线相比，RecursiveMAS平均准确率提升8.3%，端到端推理速度加快1.2-2.4倍，令牌使用量减少34.6%-75.6%。代码与数据详见https://recursivemas.github.io。

English

Recursive or looped language models have recently emerged as a new scaling axis by iteratively refining the same model computation over latent states to deepen reasoning. We extend such scaling principle from a single model to multi-agent systems, and ask: Can agent collaboration itself be scaled through recursion? To this end, we introduce RecursiveMAS, a recursive multi-agent framework that casts the entire system as a unified latent-space recursive computation. RecursiveMAS connects heterogeneous agents as a collaboration loop through the lightweight RecursiveLink module, enabling in-distribution latent thoughts generation and cross-agent latent state transfer. To optimize our framework, we develop an inner-outer loop learning algorithm for iterative whole-system co-optimization through shared gradient-based credit assignment across recursion rounds. Theoretical analyses of runtime complexity and learning dynamics establish that RecursiveMAS is more efficient than standard text-based MAS and maintains stable gradients during recursive training. Empirically, we instantiate RecursiveMAS under 4 representative agent collaboration patterns and evaluate across 9 benchmarks spanning mathematics, science, medicine, search, and code generation. In comparison with advanced single/multi-agent and recursive computation baselines, RecursiveMAS consistently delivers an average accuracy improvement of 8.3%, together with 1.2times-2.4times end-to-end inference speedup, and 34.6%-75.6% token usage reduction. Code and Data are provided in https://recursivemas.github.io.

递归多智能体系统

Recursive Multi-Agent Systems

摘要

Support