[论文] From Natural Language to Exact Cover: A Neuro-Symbolic Approach to Zeb...

论文概要

研究领域: ML 作者: Paulius Skaisgiris, Thomas Pammer, Veronika Semmelrock, Mykyta Ielanskyi, Maximilian Heisinger, Erich Kobler 发布时间: 2026-11-02 arXiv: 2611.03641

中文摘要

思维链（CoT）生成显著提升了大语言模型在复杂推理任务上的表现，包括代码生成、数据分析与考试问答。尽管如此，纯神经 LLM 在基础逻辑推理题上仍然吃力，缺乏符号推理系统所具备的确定性、可靠性与可靠性。另一方面，SAT 求解、精确覆盖（exact cover）等经典符号方法虽能保证正确与完备，但要求问题被表达为高度专门化的形式编码，难以直接处理自然语言输入。本工作提出一个神经-符号协同框架弥合这一鸿沟：结合神经语义解析与确定性精确覆盖推理。给定自然语言描述的斑马式逻辑谜题，LLM 将实体与约束翻译为结构化工具调用，构建精确覆盖实例；符号求解器计算解，并利用斑马谜题的唯一解特性对无效、欠约束或过约束的编码进行迭代反馈与修正。我们在多个参数规模、后训练范式和 LLM 家族上进行了系统的实证评估，结果表明这一混合方法持续优于包括 CoT 提示在内的强纯神经基线以及近期的神经-符号方法。

原文摘要

Chain-of-Thought (CoT) generation has substantially improved the performance of Large Language Models (LLMs) on complex reasoning tasks, including code generation, data analysis, and exam-style question answering. Despite these advances, purely neural LLMs continue to struggle with elementary logical reasoning problems and lack the determinism, soundness, and reliability characteristic of symbolic reasoning systems. Conversely, classical symbolic methods such as SAT solving and exact cover guarantee correctness and completeness, but require problems to be expressed in highly specialized formal encodings, limiting their applicability to natural language inputs. In this work, we present a synergistic neuro-symbolic framework that bridges this gap by combining neural semantic parsing with det...

#论文 #arXiv #ML #小凯