Debatepedia 和 Twelve Angry Men

本研究构建了两个自然语言论据数据集：Debatepedia和Twelve Angry Men，用于结合文本蕴含框架和双极抽象论证。Debatepedia数据集从在线辩论平台如Debatepedia和ProCon中收集，包含320个论据对，用于自动识别论据间的关系（如攻击、支持）。Twelve Angry Men数据集则从电影《十二怒汉》的剧本中提取，包含80个论据对，用于分析陪审团的讨论过程。这些数据集旨在为论证系统提供数据驱动的论证框架示例，解决自然语言论证自动处理的问题。

This study develops two natural language argumentation datasets, namely Debatepedia and Twelve Angry Men, for integrating textual entailment frameworks and bipolar abstract argumentation. The Debatepedia dataset is gathered from online debate platforms including Debatepedia and ProCon, containing 320 argument pairs, and is designed to automatically identify relational connections between arguments such as attack and support. The Twelve Angry Men dataset is extracted from the screenplay of the film *Twelve Angry Men*, consisting of 80 argument pairs, and is employed to analyze the jury's deliberation process. These datasets aim to provide data-driven examples of argumentation frameworks for argumentation systems, addressing the challenges in automatic natural language argument processing.