多智能体系统在测试时协同进化，实现个体、团队与群体的动态优化。
EVOCHAMBER: Test-Time Co-evolution of Multi-Agent System at Individual, Team, and Population Scales
- 引入CODREAM机制，团队失败后异步传递知识，保留专长差异。
- 在数学竞赛、编码和多领域推理任务上分别达63.9%、75.7%、87.1%准确率。
- 无需训练即可自发形成4-5个稳定专业分工的智能体，适合复杂协作场景。
我们提出多智能体测试时进化并非单智能体进化的简单重复。单智能体仅能演化自身上下文与记忆，而多智能体系统还能演化协作对象、协作方式及知识流动路径，产生如涌现专长等现象。但现有方法或局限于个体经验，放弃跨智能体学习；或对所有智能体对称广播，抹除协作价值。本文提出EVOCHAMBER，一个无训练框架，在个体、团队与种群三个层面实现协同进化。核心是CODREAM（协作幻梦）协议：在团队失败或分歧时，智能体协同反思、提炼洞见，并异步将知识从强到弱定向传递至失败细分领域，保留专长同时填补知识空白。团队级算子在线构建基于细分领域的团队并选择协作结构；种群级生命周期算子在性能压力下执行分裂、合并、剪枝与播种。在三个异构任务流上使用Qwen3-8B，EVOCHAMBER在竞赛数学、编码、多领域推理任务上分别达到63.9%、75.7%、87.1%准确率，相比最优基线相对提升32%（数学任务），消融实验证实异步跨智能体传递是主要驱动力。初始相同初始化的智能体中，自发涌现出4至5个稳定的专业化细分领域专家，展现出多智能体进化特有的结构性特征，单智能体无法实现。
原文摘要 · Abstract (English)
We argue that multi-agent test-time evolution is not single-agent evolution replicated N times. A single-agent learner can only evolve its own context and memory. A multi-agent system additionally evolves who collaborates, how they collaborate, and how knowledge flows across the population. These components have no single-agent counterpart and can produce phenomena such as emergent specialization. Yet prior test-time methods either confine experiences to individual agents, forfeiting cross-agent learning, or broadcast symmetrically to all agents, erasing the specialization that makes collaboration valuable. We present EVOCHAMBER, a training-free framework that instantiates test-time evolution at three levels over a coevolving agent pool. At its core is CODREAM (Collaborative Dreaming), a post-task protocol triggered on team failure or disagreement, in which agents collaboratively reflect, distill insights, and route them asymmetrically from strong to weak agents on the failed niche, preserving specialization while filling knowledge gaps. Team-level operators assemble niche-conditioned teams and select collaboration structures online. Population-level lifecycle operators fork, merge, prune, and seed agents under performance pressure. On three heterogeneous task streams with Qwen3-8B, EVOCHAMBER reaches 63.9% on competition math, 75.7% on code, and 87.1% on multi-domain reasoning, outperforming the best baseline by 32% relative on math and confirming asymmetric cross-agent transfer as the primary driver in ablation. Starting from several identically initialized agents, four to five stable niche specialists spontaneously emerge, a structural signature of multi-agent evolution that no single-agent learner can express. See our code at: https://github.com/Mercury7353/EvoChamber
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。