Scholarlink

This paper presents a large-scale biomedical Named Entity Recognition (NER) dataset automatically annotated using a Large Language Model (LLM) applied to the eligibility criteria from ClinicalTrials.gov. The dataset comprises over 4.6 million named entities, covering categories such as diseases, interventions, outcomes, and participants. A pseudo-labeling approach was employed to generate annotations with soft labels, providing confidence scores for each entity. We address challenges related to entity ambiguity and label inconsistency by introducing a structured mapping strategy to ensure uniformity across the dataset. The resulting dataset is a valuable resource for advancing tasks such as NER, information extraction, and text classification in biomedical research. By making this dataset publicly available, we aim to support the development of AI-driven healthcare applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automatic Annotation of Clinicaltrials.gov Entities Using Large Language Models

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Pouyan Nahed,
  • Sepideh Farivar,
  • Kazem Taghva

摘要

This paper presents a large-scale biomedical Named Entity Recognition (NER) dataset automatically annotated using a Large Language Model (LLM) applied to the eligibility criteria from ClinicalTrials.gov. The dataset comprises over 4.6 million named entities, covering categories such as diseases, interventions, outcomes, and participants. A pseudo-labeling approach was employed to generate annotations with soft labels, providing confidence scores for each entity. We address challenges related to entity ambiguity and label inconsistency by introducing a structured mapping strategy to ensure uniformity across the dataset. The resulting dataset is a valuable resource for advancing tasks such as NER, information extraction, and text classification in biomedical research. By making this dataset publicly available, we aim to support the development of AI-driven healthcare applications.