PEEK
收藏资源简介:
PEEK数据集是由伦敦大学学院人工智能中心创建的,包含超过20,000名非正式学习者在观看超过10,200个独特的教育视频讲座时的观看时间交互数据。该数据集通过将教育视频讲座与相关维基百科概念关联,提供了一个直观的分类体系。数据集的创建过程涉及视频的转录、翻译和片段化,以及通过Wikification技术将视频片段与维基百科概念关联。PEEK数据集的应用领域主要集中在教育推荐系统和个性化电子学习平台,旨在通过细粒度的学习者参与信号和丰富的内容表示,推动构建强大的个性化算法,从而革新教育和信息推荐系统。
The PEEK dataset was created by the Centre for Artificial Intelligence at University College London. It encompasses viewing-time interaction data from more than 20,000 informal learners who watched over 10,200 unique educational video lectures. The dataset offers an intuitive classification framework by aligning educational video lectures with relevant Wikipedia concepts. Its creation workflow involves video transcription, translation, and segmentation, as well as linking video segments to Wikipedia concepts via Wikification technology. The primary application domains of the PEEK dataset include educational recommendation systems and personalized e-learning platforms, with the objective of advancing the development of robust personalized algorithms through fine-grained learner engagement signals and rich content representations, ultimately revolutionizing educational and information recommendation systems.

- 1PEEK: A Large Dataset of Learner Engagement with Educational Videos伦敦大学学院人工智能中心 · 2021年



