Argument Summarisation and Evaluation (ASE)

Argument Summarisation and Evaluation (ASE)数据集由曼彻斯特大学创建，旨在支持自动化辩论系统的开发与评估。该数据集包含14,000个经过全面标注的声明示例，覆盖了辩论准备的全过程，包括声明和证据识别、证据说服力排名、辩论论文摘要生成及人类偏好排名等任务。数据集的创建过程涉及从在线论坛、新闻和现有数据集中收集证据，并使用大型语言模型自动生成摘要。ASE数据集的应用领域主要集中在自动化辩论辅助系统的设计与评估，旨在通过模拟人类偏好，提高辩论系统的质量和说服力。

Argument Summarisation and Evaluation (ASE) dataset was created by the University of Manchester to support the development and evaluation of automated debate systems. This dataset contains 14,000 fully annotated claim examples, covering the entire pipeline of debate preparation, including tasks such as claim and evidence identification, evidence persuasiveness ranking, debate essay summarization generation, and human preference ranking. The dataset is constructed by gathering evidence from online forums, news articles and existing datasets, and automatically generating summaries using large language models. The primary application scenarios of the ASE dataset focus on the design and evaluation of automated debate assistance systems, aiming to improve the quality and persuasiveness of debate systems by simulating human preferences.