临床主张/否定分类BERT

模型描述

在文献 Assertion Detection in Clinical Notes: Medical Language Models to the Rescue? 中引入了临床主张和否定分类BERT。该模型通过将病历中提到的医疗状况分类为存在、不存在和可能来帮助结构化临床患者病历中的信息。

该模型基于Alsentzer等人的文献 ClinicalBERT - Bio + Discharge Summary BERT Model ，并在文献 2010 i2b2 challenge 中的主张数据上进行了微调。

您可以通过transformers库加载该模型：

from transformers import AutoTokenizer, AutoModelForSequenceClassification, TextClassificationPipeline
tokenizer = AutoTokenizer.from_pretrained("bvanaken/clinical-assertion-negation-bert")
model = AutoModelForSequenceClassification.from_pretrained("bvanaken/clinical-assertion-negation-bert")

该模型的输入应以带有一个标记实体的span/句子的形式给出，以进行存在(0)、不存在(1)或可能(2)的分类。所讨论的实体用特殊标记[entity]标记。

示例输入和推断：

input = "The patient recovered during the night and now denies any [entity] shortness of breath [entity]."

classifier = TextClassificationPipeline(model=model, tokenizer=tokenizer)

classification = classifier(input)
# [{'label': 'ABSENT', 'score': 0.9842607378959656}]

引用

使用该模型时，请按照以下方式引用我们的论文：

@inproceedings{van-aken-2021-assertion,
    title = "Assertion Detection in Clinical Notes: Medical Language Models to the Rescue?",
    author = "van Aken, Betty  and
      Trajanovska, Ivana  and
      Siu, Amy  and
      Mayrdorfer, Manuel  and
      Budde, Klemens  and
      Loeser, Alexander",
    booktitle = "Proceedings of the Second Workshop on Natural Language Processing for Medical Conversations",
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.nlpmc-1.5",
    doi = "10.18653/v1/2021.nlpmc-1.5"
}