REFLECTSUMM

REFLECTSUMM是一个专为学生反思写作总结设计的新型数据集，由匹兹堡大学计算机科学系创建。该数据集包含17,512条来自24个大型STEM课程的782次大学讲座的学生反思。数据集旨在促进开发和评估针对训练数据较少但具有实际应用场景的总结技术，特别是在教育领域。REFLECTSUMM不仅包含多样化的总结任务和全面的元数据，还提供了三种类型的参考总结：提取式、抽象式和短语级提取式总结。此外，数据集还包括学生人口统计信息，有助于研究公平性和偏见问题。REFLECTSUMM的应用领域包括教育技术的改进和学生学习过程的监控。

REFLECTSUMM is a novel dataset specifically designed for student reflective writing summaries, created by the Department of Computer Science at the University of Pittsburgh. This dataset contains 17,512 student reflections from 782 university lectures across 24 large-scale STEM courses. The dataset aims to facilitate the development and evaluation of summarization techniques that perform well with limited training data but have practical real-world applications, particularly in the educational domain. REFLECTSUMM not only includes diverse summarization tasks and comprehensive metadata, but also provides three types of reference summaries: extractive, abstractive, and phrase-level extractive summaries. Additionally, the dataset contains student demographic information, which supports research on fairness and bias issues. The application areas of REFLECTSUMM include educational technology improvement and student learning process monitoring.

搜集汇总

背景与挑战

背景概述

REFLECTSUMM是一个专注于学生反思写作总结的数据集，由匹兹堡大学计算机科学系创建，包含来自24个STEM课程的17,512条学生反思，旨在支持教育领域总结技术的开发和评估。该数据集提供多样化的总结任务、三种参考总结类型以及学生人口统计信息，有助于研究公平性和偏见问题，应用领域包括教育技术改进和学生学习监控。