遞歸多智能體系統

摘要

近期，遞歸式或循環語言模型通過對潛在狀態進行迭代式模型計算以深化推理，已成為新的規模化拓展軸向。我們將此規模化原則從單一模型擴展至多智能體系統，並提出關鍵問題：智能體協作本身能否通過遞歸實現規模化？為此，我們提出RecursiveMAS——一個將整個系統視為統一潛在空間遞歸計算的遞歸多智能體框架。該框架通過輕量級RecursiveLink模塊將異構智能體連接為協作循環，實現分佈內潛在思維生成與跨智能體潛在狀態傳輸。為優化框架，我們開發了內外雙環學習算法，通過遞歸輪次間的基於梯度的共享信用分配，實現迭代式全系統協同優化。對運行時間複雜度與學習動態的理論分析表明，RecursiveMAS較標準基於文本的多智能體系統更具效率，並能在遞歸訓練中保持梯度穩定性。實證研究中，我們在4種代表性智能體協作模式下實例化RecursiveMAS，並在涵蓋數學、科學、醫學、搜索及代碼生成的9個基準測試中進行評估。相較先進的單一/多智能體及遞歸計算基線，RecursiveMAS持續實現8.3%的平均準確率提升，同時帶來1.2倍至2.4倍的端到端推理加速，以及34.6%-75.6%的標記使用量削減。代碼與數據已發佈於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.

遞歸多智能體系統

Recursive Multi-Agent Systems

摘要

Support