Essential Step Detection (ESD)

ESD数据集由宾夕法尼亚大学和清华大学联合创建，专注于事件过程的必要性检测。该数据集包含1515对目标-步骤数据，从WikiHow网站收集，并由专家手动标注步骤的必要性。数据集旨在评估模型对事件过程中关键步骤的理解能力，应用于提升自然语言处理中的事件理解和常识推理能力。

The ESD dataset was co-created by the University of Pennsylvania and Tsinghua University, focusing on necessity detection in event processes. It contains 1515 target-step pairs collected from the WikiHow website and manually annotated for the necessity of each step by domain experts. This dataset is designed to evaluate models' ability to understand the critical steps within event processes, and is applied to enhance event understanding and commonsense reasoning capabilities in natural language processing (NLP).

搜集汇总

背景与挑战

背景概述

Essential Step Detection (ESD)数据集由宾夕法尼亚大学和清华大学联合创建，从WikiHow网站收集1515对目标-步骤数据，并由专家手动标注步骤的必要性。该数据集专注于事件过程中的必要性检测，旨在评估模型对关键步骤的理解能力，以提升自然语言处理中的事件理解和常识推理能力。