An Enclosed Mode Is a Gauge Choice: Topology Relative to Reach in Certified Code World Models

📄 arXiv: 2608.28541v1 📥 PDF

作者: Javier Aguilar Martín

分类: cs.LG, cs.AI

发布日期: 2026-08-28

备注: 33 pages, 2 figures. Paper 3 of a series (companion papers: arXiv:2607.14169, arXiv:2608.17956). Code, data, and Lean formalization: https://github.com/JaviMaligno/code-world-models


💡 一句话要点

提出一种新方法以解决认证代码世界模型中的拓扑问题

🎯 匹配领域: 支柱二:RL算法与架构 (RL & Architecture)

关键词: 认证模型 拓扑分析 采样门 环形仪器 错误成本 持久同调 智能系统 决策能力

📋 核心要点

  1. 现有的认证模型在可达范围内表现良好,但在超出可达范围时可能出现任意错误,导致模型的不可靠性。
  2. 论文提出了一种通过采样门和环形仪器来精确界定模型知识和错误成本的新方法,强调拓扑与可达性之间的关系。
  3. 实验结果表明,通过调整通道宽度,可以显著改变模型的表现,验证了三条组织实验的原则,提供了新的见解。

📝 摘要(中文)

本文探讨了在采样门接受的代码世界模型中,模型在可达查询集上的准确性与超出可达范围的任意错误之间的关系。我们定义了认证模型的知识范围及其错误成本,特别是在封闭的不可达内部的环形冻结模式下。通过对最小环形仪器的极端案例证明,我们展示了如何通过调整通道宽度来影响模型的表现,并提出了三条原则来组织实验结果,强调了拓扑相对于可达性的危险性、修复的参数和传感器限制,以及缓解措施必须与错误的维度和方向相匹配。

🔬 方法详解

问题定义:本文旨在解决认证代码世界模型在可达范围外的错误问题,现有方法未能有效处理拓扑与可达性之间的关系。

核心思路:通过引入采样门和环形仪器,精确界定模型的知识范围及其错误成本,特别是在不可达区域的表现。

技术框架:整体架构包括采样门的设计、环形仪器的构建,以及对不同通道宽度下模型表现的评估。主要模块包括模型知识界定、错误成本分析和实验验证。

关键创新:提出了基于拓扑的危险性评估和修复策略,强调了参数和传感器的限制,提供了与现有方法的本质区别。

关键设计:在实验中,通道宽度的设置为gamma,损失函数设计考虑了拓扑错误的维度,网络结构采用了持久同调的总结方法来跟踪错误参数。

🖼️ 关键图片

img_0
img_1
img_2

📊 实验亮点

实验结果显示,通过调整通道宽度gamma,模型的表现从1.09降至接近0,验证了拓扑相对于可达性的危险性。同时,采用持久同调方法的模型在错误参数追踪上表现出显著的改进,提升幅度达到1.769到0.029。

🎯 应用场景

该研究在机器人、自动驾驶和智能系统等领域具有广泛的应用潜力。通过改进模型的可靠性和准确性,可以提升系统在复杂环境中的决策能力,推动智能系统的安全性和效率。

📄 摘要(原文)

A code world model accepted by a sampling gate can be exactly right on everything the gate can see and arbitrarily wrong beyond it. We characterize what a certified model can know, and what its errors can cost, when the omission is an annular freeze mode enclosing an unreachable interior. The gate quotient makes the question precise: acceptance-with-certainty determines the model exactly on the reachable query set; beyond reach is gauge. On a minimal ring instrument we prove the extreme case (a wrong-topology filled-disc artifact unfalsifiable by any sampling gate and bitwise harmless at play) and measure, with LLM synthesis across three model families, how one knob (a channel of width gamma) walks the same artifact through three regimes: unfalsifiable-and-harmless, falsifiable-and-costly, and instantly falsified. Three principles organize the empirics. First, danger is topology relative to reach: a channel the planner can use collapses the blind model's exploitation (play cost 1.09 to ~0 over a knee at gamma ~ 0.1), while a hidden channel with the same first Betti number keeps it at full strength (1.12). Second, repair is parameter-bound and sensor-bound: no family recovers the region from outside evidence; from inside, models pose the right topology but cannot pin its parameters, and the posed topology tracks the guiding persistent-homology summary's wrong beta_1 (a sensor with a measured geometric resolution limit), not the truth. Third, mitigation must match the error's dimension and direction: point fences fail against the one-dimensional boundary, a dimension-matched persisted fence collapses exploitation to a two-lesson transient (0.999 to 0.058), and the dual freedom certificate collapses the invented-mode failure symmetrically (1.769 to 0.029). In n dimensions the shell makes misidentification near-certain while the danger stays fully exploitable: the two axes are independent.