智能体社会是否发展出知识分子?LLM多智能体系统中集体认知的隐藏幂律
Do Agent Societies Develop Intellectual Elites? The Hidden Power Laws of Collective Cognition in LLM Multi-Agent Systems
AI总结:
研究LLM多智能体系统中协调动态,揭示协调遵循幂律级联、集中于知识分子精英及系统规模增长时极端事件频发的三大定律,提出整合瓶颈机制及Deficit-Triggered Integration方法提升性能。
摘要:
大型语言模型(LLM)多智能体系统日益被部署为交互智能体社会,但扩展这些系统往往导致收益递减或不稳定,其原因尚不明确。本文首次对LLM多智能体系统的协调动态进行大规模实证研究,引入原子事件级公式重构推理为级联协调。分析超过150万次交互,发现三大耦合定律:协调遵循重尾级联,通过偏好附件集中于知识分子精英,系统规模增长时极端事件频发。证明这些效应通过单一结构机制耦合:整合瓶颈,协调扩展随系统规模增长,而整合不增长,产生大但弱整合的推理过程。为测试此机制,引入Deficit-Triggered Integration(DTI),在不平衡时选择性增加整合。DTI在协调失败处提升性能,不抑制大规模推理。结果确立集体认知的定量定律,识别协调结构为理解及改进可扩展多智能体智能的关键轴。
英文摘要:
Large Language Model (LLM) multi-agent systems are increasingly deployed as interacting agent societies, yet scaling these systems often yields diminishing or unstable returns, the causes of which remain poorly understood. We present the first large-scale empirical study of coordination dynamics in LLM-based multi-agent systems, introducing an atomic event-level formulation that reconstructs reasoning as cascades of coordination. Analyzing over 1.5 Million interactions across tasks, topologies, and scales, we uncover three coupled laws: coordination follows heavy-tailed cascades, concentrates via preferential attachment into intellectual elites, and produces increasingly frequent extreme events as system size grows. We show that these effects are coupled through a single structural mechanism: an integration bottleneck, in which coordination expansion scales with system size while consolidation does not, producing large but weakly integrated reasoning processes. To test this mechanism, we introduce Deficit-Triggered Integration (DTI), which selectively increases integration under imbalance. DTI improves performance precisely where coordination fails, without suppressing large-scale reasoning. Together, our results establish quantitative laws of collective cognition and identify coordination structure as a fundamental, previously unmeasured axis for understanding and improving scalable multi-agent intelligence.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。