Automation of Trainable Datasets Generation for Medical-Specific Language Model: Using MIMIC-IV Discharge Notes
Studies in Health Technology and Informatics · 2024
DOI: 10.3233/shti240497 ↗Kurz gefasst
Fachartikel in der Zeitschrift Studies in Health Technology and Informatics, 2024. Verfasst von Youngrong Lee und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Medical Informatics, College of Medicine, The Catholic University of Korea, Republic of Korea. Hauptthemen laut MEDLINE: Natural Language Processing, Electronic Health Records. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
This study introduces a novel approach for generating machine-generated instruction datasets for fine-tuning medical-specialized language models using MIMIC-IV discharge records. The study created a large-scale text dataset comprising instructions, cropped discharge notes as inputs, and outputs in JSONL format. The dataset was generated through three main stages, generating instruction and output using seed tasks provided by medical experts, followed by invalid data filtering. The generated dataset consisted of 51,385 sets, with mean ROUGE between seed tasks of 0.185. Evaluation of the generated dataset were promising, with high validity rates determined by both GPT-3.5 and a human annotator (88.0% and 88.5% respectively). The study highlights the potential of automating dataset creation for NLP tasks in the medical domain.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Medical Informatics, College of Medicine, The Catholic University of Korea, Republic of Korea.
- Department of Biomedicine & Health Sciences, College of Medicine, The Catholic University of Korea, Republic of Korea.
Zitieren
APA
Lee, Y., Kim, C., & Ko, T. (2024). Automation of Trainable Datasets Generation for Medical-Specific Language Model: Using MIMIC-IV Discharge Notes. Studies in Health Technology and Informatics. https://doi.org/10.3233/shti240497
Vancouver
Lee Y, Kim C, Ko T. Automation of Trainable Datasets Generation for Medical-Specific Language Model: Using MIMIC-IV Discharge Notes. Studies in Health Technology and Informatics. 2024. doi:10.3233/shti240497
BibTeX
@article{Lee2024,
author = {Lee, Youngrong and Kim, Chansik and Ko, Taehoon},
title = {Automation of Trainable Datasets Generation for Medical-Specific Language Model: Using MIMIC-IV Discharge Notes},
journal = {Studies in Health Technology and Informatics},
year = {2024},
doi = {10.3233/shti240497}
}
EUVIMED