数据集:
bigbio/bc7_litcovid
The training and development datasets contain the publicly-available text of over 30 thousand COVID-19-related articles and their metadata (e.g., title, abstract, journal). Articles in both datasets have been manually reviewed and articles annotated by in-house models.
@inproceedings{chen2021overview, title = { Overview of the BioCreative VII LitCovid Track: multi-label topic classification for COVID-19 literature annotation }, author = { Chen, Qingyu and Allot, Alexis and Leaman, Robert and Do{\u{g}}an, Rezarta Islamaj and Lu, Zhiyong }, year = 2021, booktitle = {Proceedings of the seventh BioCreative challenge evaluation workshop} }