数据集:
bigbio/biosses
BIOSSES 通过利用 WordNet 作为一般领域本体论和 UMLS 作为生物医学领域特定本体论来计算生物医学句子的相似性。原始论文概述了根据使用注释者评分作为黄金标准的方法。源视图将分别返回所有注释者评分,而大生物视图将返回注释者评分的平均值。
@article{souganciouglu2017biosses,
title={BIOSSES: a semantic sentence similarity estimation system for the biomedical domain},
author={Soğancıoğlu, Gizem, Hakime Öztürk, and Arzucan Özgür},
journal={Bioinformatics},
volume={33},
number={14},
pages={i49--i58},
year={2017},
publisher={Oxford University Press}
}