数据集:
bigbio/ebm_pico
该语料库包含4,993个带有(P)参与者、(I)干预和(O)结果注释的摘要。训练标签来自AMT工人,并进行了聚合以减少噪音。测试标签由医疗专业人员收集。
@inproceedings{nye-etal-2018-corpus,
title = "A Corpus with Multi-Level Annotations of Patients, Interventions and Outcomes to Support Language Processing for Medical Literature",
author = "Nye, Benjamin and
Li, Junyi Jessy and
Patel, Roma and
Yang, Yinfei and
Marshall, Iain and
Nenkova, Ani and
Wallace, Byron",
booktitle = "Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2018",
address = "Melbourne, Australia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/P18-1019",
doi = "10.18653/v1/P18-1019",
pages = "197--207",
}