用循环路由让多个AI代理更智能协作，省计算还高效
MMoA: An AI-Agent framework with recurrence for Memoried Mixure-of-Agent
- 引入循环神经网络作为路由，动态调整代理参与度
- 在多个评测中准确率接近传统方法，但运行效率提升4.6%
- 适合需要高效多代理协同的大型语言模型应用
混合代理（MoA）框架通过聚合多个代理输出提升了大语言模型性能，但现有系统依赖静态路由，未能充分捕捉跨层的时间与上下文依赖。为此，我们提出MMoA，一种将基于LSTM的门控机制融入代理选择过程的递归MoA架构。该递归路由根据当前输入和历史路由决策自适应调节代理贡献，实现更上下文感知的聚合。我们在AlpacaEval 2.0、MT-Bench和Arena-Hard等标准指令遵循基准上评估了MMoA。结果表明，MMoA在保持与传统MoA相当准确率的同时，通过动态激活更少代理降低了计算开销。例如，在AlpacaEval 2.0上，MMoA取得58.0%胜率，略低于MoA的59.8%，但运行效率最高提升4.6%。这些结果表明，MMoA为自适应多代理大语言模型系统提供了一种可扩展且高效的方案。
原文摘要 · Abstract (English)
The Mixture-of-Agents (MoA) framework has shown promise in improving large language model (LLM) performance by aggregating outputs from multiple agents. However, existing MoA systems often rely on static routers that do not fully capture temporal and contextual dependencies across aggregation layers. To address this limitation, we propose MMoA, a recurrent MoA architecture that integrates LSTM-based gating into the agent selection process. The recurrence router adaptively modulates agent contributions based on both current inputs and historical routing decisions, enabling more context-aware aggregation. We evaluate MMoA on standard instruction-following benchmarks, including AlpacaEval 2.0, MT-Bench, and Arena-Hard. The results show that MMoA achieves comparable accuracy to traditional MoA while reducing computational overhead by dynamically activating fewer agents. For example, on AlpacaEval 2.0, MMoA achieves a win rate of 58.0%, compared with 59.8% for MoA, while improving runtime efficiency by up to 4.6%. These results suggest that MMoA provides a scalable and efficient approach for adaptive multi-agent LLM systems.
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。