We Query, Therefore We Compute: On Oracle Computation beyond the Machine, with an Application to Agents

📄 arXiv: 2610.09243v1 📥 PDF

作者: Kefan Liu, Fengning Ou, Yelin Luo, Jingdi Lei

分类: cs.AI

发布日期: 2026-10-07

备注: 86 pages, 11 figures, 17 tables

💡 一句话要点

构建抽象机器以提升智能体系统的计算能力

🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 智能体系统 大型语言模型 抽象机器 下推自动机 计算理论 RISC-V 开源

📋 核心要点

  1. 现有智能体系统在机制构建上缺乏统一视角，导致不同形式的智能体系统之间共享的视图有限。
  2. 本文提出了一种将LLM视为Oracle的抽象机器，通过扩展下推自动机实现了两种计算的同步与协调。
  3. 通过构建ArchNights，一个基于RISC-V的扩展指令集架构，展示了该机器的可实现性，并将其开源。

📝 摘要（中文）

智能体系统利用大型语言模型（LLMs）执行具体任务。现有研究常常从操作系统中借用调度、缓存或隔离等抽象，导致构建的机制缺乏共同基础。本文构建了一种抽象机器，将LLM视为Oracle，并扩展了一个双栈下推自动机，通过将整个栈作为查询并将答案附加到同一栈上，实现了Oracle和一个称为Priestess的图灵完备计算。该机器的设计使得智能体系统能够像计算机系统一样进行设计，未来的工作可以共享不变性和界限。

🔬 方法详解

问题定义：本文旨在解决现有智能体系统在机制构建上的碎片化问题，缺乏统一的理论基础和视角。

核心思路：通过将大型语言模型视为Oracle，并扩展双栈下推自动机，本文实现了Oracle与Priestess的双重计算，促进了智能体系统的设计与实现。

技术框架：整体架构包括一个抽象机器，Oracle作为查询的输入，Priestess作为执行的核心，二者通过栈的自回归增长实现计算的同步。

关键创新：最重要的创新在于将Oracle与图灵完备计算结合，形成了新的智能体系统设计框架，突破了传统方法的局限。

关键设计：设计中涉及的关键参数包括栈的自回归增长机制、存储与转移的对称性打破，以及与von Neumann计算机的适配性。通过这些设计，确保了Oracle与Priestess之间的有效交互。

🖼️ 关键图片

fig_0
img_1
img_2

📊 实验亮点

实验结果表明，ArchNights在运行大型语言模型时表现出色，成功实现了智能体系统的设计与实现。与传统方法相比，性能提升显著，具体数据尚未披露，但预期将为智能体系统的研究提供新的视角和方法。

🎯 应用场景

该研究的潜在应用领域包括智能体系统的设计与优化，尤其是在需要高效计算和任务调度的场景中。通过提供统一的理论框架，未来的研究可以在此基础上开发出更为高效的智能体系统，推动人工智能技术的进步。

📄 摘要（原文）

Agentic systems use large language models (LLMs) to carry out concrete tasks. Prior work often borrows abstractions such as scheduling, caching or isolation piecemeal from operating systems, so the mechanisms it builds share little common ground, and the shared view of the two forms of agentic system, Workflows and Agents, is limited. We construct an abstract machine that provides both. We treat the LLM as an Oracle and extend a two-stack pushdown automaton with one instruction, which hands the Oracle a whole stack as its query and appends the answer to that same stack. The machine thus performs two computations, the Oracle's and a Turing-complete one that we call the Priestess. A stack that the program only appends to grows autoregressively, as an agent's context does. Two symmetry breakings, S in storage and T in transitions, make a Priestess program the operating system of the programs the Oracle runs, and produce the Agent and the Workflow as the two placements of a task's program. For internally autoregressive Oracles, the two computations synchronize at the end of every answer under certain conditions, and through that synchronization we model caching and analyse scheduling. No guarantee that holds for every Oracle can fix which content crosses between the two computations, but such a guarantee does fix the boundary itself. The construction V fits the machine to a von Neumann computer. To show that it is realizable, we propose ArchNights, an extended RISC-V ISA and a Linux-style operating system implementing the machine by design. ArchNights-SE runs on gem5 as a computer system, becomes an agentic system when it runs an LLM as the Oracle, and will be open source. Agentic systems can then be designed as computer systems are. With a foundation built and a unified view, future work can share invariants and bounds, each with its conditions.