Think Through a Bottleneck: Hourglass Reasoning for Rigorous Induction
作者: Huan Zhu
分类: cs.AI
发布日期: 2026-07-13
💡 一句话要点
提出Hourglass推理以解决少样本归纳推理问题
🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)
关键词: 少样本学习 归纳推理 上下文隔离 符号推理 大型语言模型
📋 核心要点
- 现有方法在少样本归纳推理中存在自我精炼效果不佳的问题，提示模型明确推导规则的效果有限。
- 本文提出Hourglass推理，通过在推理阶段之间强制隔离上下文，确保信息以压缩符号状态传递，从而提高推理效果。
- 在多个基准测试中，Hourglass推理显著提升了模型的准确性，例如在ChipBench上将Verilog合成准确率从31%提高至58%。
📝 摘要（中文）
自我精炼在大型语言模型中的少样本归纳推理中常常失败。单纯提示模型明确其推导规则效果有限，关键在于在推理阶段之间强制隔离结构，以便信息仅能以压缩的符号状态在阶段间传递。本文提出了Hourglass推理，严格执行推理阶段之间的上下文隔离。通过将冻结的LLM作为元构造器，为每个任务构建符号编码器-解码器，显著提升了在视觉抽象、硬件合成和文本规则归纳等基准上的表现。
🔬 方法详解
问题定义：本文旨在解决大型语言模型在少样本归纳推理中的自我精炼效果不佳的问题。现有方法往往未能有效隔离推理阶段，导致信息流动不畅。
核心思路：Hourglass推理的核心在于严格的上下文隔离，确保信息仅在推理阶段之间以压缩的符号状态传递，从而增强推理的有效性。
技术框架：该方法包括多个模块：归纳模块将支持示例压缩为符号架构φ和临时支架z；推导模块从中推导规则T并丢弃z；实现者将(φ, T)编译为工件；错误驱动的精炼者修订(φ, T)并从头生成工件。
关键创新：Hourglass推理的创新在于其阶段之间的严格隔离设计，这与现有方法的直接信息传递方式形成鲜明对比，确保了推理过程的稳定性和准确性。
关键设计：在设计中，模型的初始归纳质量和阶段隔离的有效性是关键因素，实验表明这些因素对性能提升至关重要，而提示的措辞和符号形式对结果的影响较小。
🖼️ 关键图片
📊 实验亮点
在实验中，Hourglass推理在ARC-AGI-2基准上将最佳准确率提高了14个百分点，在ChipBench上将Verilog合成准确率从31%几乎翻倍至58%。在BBEH-Linguini测试中，Hourglass推理有效逆转了以往显式表述对性能的负面影响。
🎯 应用场景
Hourglass推理的潜在应用领域包括视觉抽象、硬件设计和自然语言处理等。其方法论可以为复杂推理任务提供更为稳健的解决方案，未来可能在智能系统、自动化设计和教育等领域产生深远影响。
📄 摘要（原文）
Self-refinement often fails to strengthen few-shot inductive reasoning in large language models. Prompting a model to explicitly state its inferred rule does little on its own. What actually matters is a structurally enforced isolation between reasoning stages, so that information can only pass between them as a compressed symbolic state. We introduce \textbf{Hourglass reasoning}, which enforces strict context isolation between reasoning stages. The frozen LLM acts as a meta-constructor, building for each task a symbolic encoder--decoder: an Induction module compresses the support examples into a schema $φ$ (encoder) and a transient scaffold $z$; a Deduction module derives rule $T$ (decoder) from these and discards $z$; an Implementer compiles $(φ, T)$ into artifacts; an error-driven Refiner revises $(φ, T)$ and regenerates artifacts from scratch. Only $(φ, T)$ crosses stage boundaries, so all refinement stays anchored to the rule. We evaluate Hourglass across three benchmarks spanning visual abstraction, hardware synthesis, and textual rule induction, using GPT-5.5 and Gemini 3.1 Pro. On ARC-AGI-2, it raises best-of-5 accuracy by up to 14 points over an iterative-refinement baseline. On ChipBench, it nearly doubles Verilog synthesis accuracy with GPT-5.5, from 31\% to 58\%. BBEH-Linguini draws on puzzles from the International Linguistics Olympiad, a setting where prior work has shown that explicit verbalization can hurt performance. Hourglass mitigates this tendency, and on Gemini 3.1 Pro, it reverses the effect entirely. Ablations confirm that these gains come from the isolation between stages and the quality of the initial induction, not from prompt wording or the particular symbolic form used. It is how information flows through the reasoning process, rather than the language used to express it, that drives inductive reasoning in frozen LLMs.