MyST
收藏资源简介:
MyST数据集是由My Science Tutor项目开发,包含约400小时的儿童对话语音,涵盖230K条语音记录,涉及1.3K名三至五年级学生与虚拟科学导师的互动。数据集内容丰富,包括8个科学领域的对话,旨在通过自动语音识别技术提升教育领域的应用。创建过程中,学生与虚拟导师进行一对一互动,记录并转录对话,确保数据匿名化。该数据集的应用领域主要集中在教育技术,特别是远程教育和科学教育,旨在提高儿童对科学的兴趣和学习效果。
The MyST Dataset was developed by the My Science Tutor project. It contains approximately 400 hours of children’s conversational speech, comprising 230,000 speech recordings, and documents interactions between 1,300 students in grades 3 to 5 and a virtual science tutor. The dataset features rich content covering conversations across 8 scientific disciplines, and is designed to facilitate the application of automatic speech recognition (ASR) technologies in the education field. During the dataset construction process, students participated in one-on-one interactions with the virtual tutor, with their dialogues recorded and transcribed, and strict anonymization procedures were applied to ensure data privacy. The main application scenarios of this dataset lie in educational technology, especially distance education and science education, with the objective of boosting children’s interest in science and improving their learning effectiveness.



