MIT Level 6 Fine-Grained, Full-Coverage, Lossless Critical Thinking Knowledge Graph

This dataset, based on three transcripts from Modules 1–3 of MIT's Critical Thinking and Decision Making course, constructs a six-level, fine-grained, quantifiable knowledge graph. It organises all key content hierarchically as ‘Course → Module → Topic → Concept → Proposition → Exemplification’, presenting argument chains and case evidence through an auditable node-edge structure. The dataset simultaneously provides structural edges (contains) and semantic edges (exemplifies). Quantitative information such as numbers, percentages, year ranges, and ratios is extracted and structured within nodes. This supports diverse research and engineering applications including GraphRAG retrieval, argument analysis, course content mining, and decision modelling.

本数据集基于麻省理工学院（MIT）《批判性思维与决策》课程第1至3模块的三份课程转录文本，构建了一套六层级、细粒度且可量化的知识图谱（Knowledge Graph）。该数据集将所有核心内容按层级结构组织为「课程→模块→主题→概念→命题→示例」的架构，并通过可审计的节点-边结构呈现论证链条与案例佐证。本数据集同时提供两类边：结构边（包含关系）与语义边（示例关系）。包括数值、百分比、年份区间与比例在内的各类量化信息均被提取并结构化存储于节点中。本数据集可支撑多类研究与工程应用，涵盖GraphRAG检索、论证分析、课程内容挖掘与决策建模。