[论文] Shepherd: A Runtime Substrate Empowering Meta-Agents with a Formalized...

论文概要

研究领域: ML 作者: Simon Yu, Derek Chong, Ananjan Nandi 发布时间: 2025-05-09 arXiv: 2505.07236

中文摘要

我们介绍了Shepherd，一种函数式编程模型，将元智能体对目标智能体的操作形式化为函数，核心操作在Lean中机械化实现。Shepherd将每个智能体-环境交互记录为Git式执行跟踪中的类型化事件，使任何过去的状态都可以被分叉和重放。该系统分叉智能体进程及其文件系统的速度比Docker快5倍，重放时实现>95%的提示缓存复用。我们通过三个应用展示了该模型...

原文摘要

We introduce Shepherd, a functional programming model that formalizes meta-agent operations on target agents as functions, with core operations mechanized in Lean. Shepherd records every agent-environment interaction as a typed event in a Git-like execution trace, enabling any past state to be forked and replayed. The system forks the agent process and its filesystem 5x faster than Docker, achieving >95% prompt-cache reuse on replay. We demonstrate the model through three applications...

#论文 #arXiv #ML #小凯