CERA-MoA: Co-Evolving Routing Mechanisms with Continually Learning LLM Agents· CERA-MoA: 持续学习的 LLM 代理的共进化路由机制
Current Mixture-of-Agents (MoA) paradigms generally treat query routing and agent fine-tuning as separate processes, limiting their ability to respond to evolving agent capabilities. This disconnect prevents routing strategies from adapting to evolving agent capabilities during post-training and prevents agents from achieving synergistic data-driven specialization. To resolve this, we introduce CERA-MoA (Co-Evolving Router with continually learning Agents for Mixture-of-Agents), an iterative reinforcement learning framework where the dynamic router and independent agent policies co-evolve. We design a predictive familiarity estimator that leverages mid-layer hidden states to evaluate semantic competence among agents, avoiding the overhead of full rollouts. Based on these familiarity scores
提出共进化的路由机制与持续学习的LLM代理,增强MoA系统的适应性与效率。
- 核心方法
- 设计CERA-MoA框架,通过迭代强化学习实现动态路由机制与独立代理政策的共进化,并引入预测熟悉度估计器利用中间层隐藏状态评估代理的语义能力。
- 适合谁读
- 研究者
- 要解决的问题
- 现有MoA框架中的查询路由与代理的微调过程分离,导致路由策略无法适应代理能力的变化,影响系统整体性能。
- 关键实验
- 未提供具体实验数据,但描述了框架在理论上能够有效适应代理能力的变化。
- 主要贡献
- 解决了路由策略与代理能力不匹配的问题,提高了系统的适应性和代理的协作效率。
- 意义与局限
- 为多代理系统提供了一种新的动态优化方法,有望提高大型语言模型(LLM)在实际应用中的灵活性和性能。局限性在于理论部分较强,缺乏实际验证。