PulseAugur
实时 15:44:25
English(EN) Training Small LLMs as Spatial Multi-Agent Policies

新研究探索使用符号选项和协同进化策略训练的LLM智能体

研究人员正在探索训练大型语言模型(LLM)作为复杂环境中智能体的新方法。其中一种方法,在一篇新的arXiv论文中有详细介绍,使用“前沿编码模型”为智能体生成符号选项,然后使用多智能体GRPO(PA-MAGRPO)的变体和LoRA适配器进行训练。该方法旨在通过关注行为评估来提高智能体性能,超越简单的奖励最大化,因为奖励本身可能无法可靠地指示真正的合作。另一篇论文介绍了一种用于资源受限LLM智能体训练后优化的协同参数子空间进化策略(CoPES),证明与标准的进化策略和GRPO相比,它可以在显著降低内存需求的情况下实现性能的大幅提升。 AI

影响 这些方法可以实现更复杂、更高效的LLM智能体训练,尤其是在资源受限的环境中,有可能带来更强大的AI系统。

排序理由 该集群包含两篇学术论文,详细介绍了训练LLM智能体的新颖方法。

在 arXiv cs.LG 阅读 →

AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →

新研究探索使用符号选项和协同进化策略训练的LLM智能体

报道来源 [3]

  1. arXiv cs.LG TIER_1 English(EN) · Yi Mao, Andrew Perrault ·

    Training Small LLMs as Spatial Multi-Agent Policies

    arXiv:2608.01425v1 Announce Type: cross Abstract: Training LLM-based multi-agent systems with multi-agent reinforcement learning is rapidly gaining traction, and a parallel line of work argues that such systems should be judged by their behavior, not only their reward. We take up…

  2. arXiv cs.LG TIER_1 English(EN) · Zhiyuan Wang, Shengcai Liu, Jiahao Wu, Ning Lu, Hui Ouyang, Shaofeng Zhang, Haoze Lv, Ke Tang ·

    Cooperative Coevolution for Resource-Constrained Agentic LLM Post-Training

    arXiv:2608.02391v1 Announce Type: cross Abstract: Tool-using large language model (LLM) agents produce long, multi-turn trajectories, making gradient-based post-training memory-intensive. Evolution strategies (ES) enable memory-efficient full-parameter post-training without backp…

  3. arXiv cs.MA (Multiagent) TIER_1 English(EN) · Andrew Perrault ·

    Training Small LLMs as Spatial Multi-Agent Policies

    Training LLM-based multi-agent systems with multi-agent reinforcement learning is rapidly gaining traction, and a parallel line of work argues that such systems should be judged by their behavior, not only their reward. We take up both threads in spatial cooperative games, where …