From Events to Reasoning: Hierarchical Rule Composition over Explicit States

Computer Science > Artificial Intelligence arXiv:2610.22403 (cs) [Submitted on 14 Oct 2026] Title: From Events to Reasoning: Hierarchical Rule Composition over Explicit States Authors: Cheng Lyu, Xueqi Wang.

计算机科学 > 人工智能 arXiv:2610.22403 (cs) [提交于 2026 年 10 月 14 日] 标题：从事件到推理：显式状态下的分层规则组合 作者：Cheng Lyu, Xueqi Wang。

Abstract: What a reasoning system emits is a conclusion; what it owes is a justification. Producing one requires every operation to be admissible on the facts actually available, the dependencies of intermediate results to survive to the end of the path, and the path to be revised when a premise changes.

摘要：推理系统输出的是结论，而它所欠缺的是论证。要产生一个结论，要求每一次操作都必须基于实际可用的事实，中间结果的依赖关系必须在路径终点依然有效，并且当前提发生变化时，路径必须能够进行修订。

We present the Event Reasoning System (ERS), in which the reasoning state is an explicit event graph. Semantic features nominate candidate objects and rule bindings, and declared rule and task conditions decide which may execute, so candidates that violate the represented period, unit, or evidence constraints cannot enter an accepted reasoning path.

我们提出了事件推理系统（ERS），其中推理状态是一个显式的事件图。语义特征提名候选对象和规则绑定，声明的规则和任务条件决定哪些可以执行，因此违反所表示的时间段、单位或证据约束的候选者无法进入被接受的推理路径。

Low-level rules provide reusable local transformations, and recursively composable high-level control rules organize subgoals and legal paths under the current goal, information gaps, and budget; a finite library is thereby rebound into paths never stored as complete solutions.

低级规则提供可重用的局部转换，而可递归组合的高级控制规则则在当前目标、信息缺口和预算下组织子目标和合法路径；有限的库因此被重新组合成从未作为完整解决方案存储的路径。

We formalize admissible transitions, task-dependent control, compositional expressivity, dependency-preserving revision, and the gap between reachable solutions and those a budgeted controller finds.

我们对可接受的转换、任务相关的控制、组合表达能力、依赖保持修订，以及可达解与预算控制器所找到的解之间的差距进行了形式化。

On structured-event graphs, ERS builds dependency chains of up to twelve steps, cuts irrelevant executed domain operations from 32 to 0 as lookahead grows to 16, and produces no unsupported target output across 24 negative cases; with rules frozen, it passes 12/12 focused checks comprising nine accepted derivations involving unseen input/path combinations and three justified refusals, as well as 6/6 fact-change cases.

在结构化事件图上，ERS 构建了长达十二步的依赖链，随着前瞻深度增加到 16，将不相关的已执行领域操作从 32 次减少到 0 次，并且在 24 个负面案例中没有产生任何不受支持的目标输出；在规则固定的情况下，它通过了 12/12 的重点检查，包括九个涉及未见输入/路径组合的接受推导和三个合理的拒绝，以及 6/6 的事实变更案例。

On fourteen jointly solved development tasks, the recorded dynamic subgoal configuration uses 51.6% less inference wall-clock time than single-level global control under their respective frozen settings.

在十四个联合解决的开发任务中，记录的动态子目标配置比各自固定设置下的单级全局控制减少了 51.6% 的推理挂钟时间。