Chinese electronic medical record named entity recognition model based on pre-training and multi-task learning

Jiakang Li; Ruixia Liu; Lihui Su; Shikai Zhang

doi:10.1117/12.2640089

1 August 2022 Chinese electronic medical record named entity recognition model based on pre-training and multi-task learning

Jiakang Li, Ruixia Liu, Lihui Su, Shikai Zhang

Author Affiliations +

Proceedings Volume 12257, 4th International Conference on Information Science, Electrical, and Automation Engineering (ISEAE 2022); 122570R (2022) https://doi.org/10.1117/12.2640089
Event: 4th International Conference on Information Science, Electrical, and Automation Engineering (ISEAE 2022), 2022, Guangzhou, China

Abstract

The Chinese Electronic Medical Records (EMR) lacks abundant annotated data. This severely limits the performance of Named Entity Recognition (NER) models in this domain. We propose a Chinese electronic medical record named entity recognition model based on pre-training and multi-task learning (Pt-Mt) to solve this problem. The model first fine-tunes the improved pre-trained model Roberta on different medical datasets, so that Roberta better fits the characteristics of the medical field and can learn features on different datasets. At the same time, the Chinese Word Segmentation (CWS) task is added as an auxiliary task of the NER model for joint training, which enhances the model's ability to distinguish entity boundaries. Finally, the NER of Chinese EMR based on Roberta and multi-task learning has achieved good results on the CCKS2017, CCKS2019, and CCKS2020.

Citation Download Citation

Jiakang Li, Ruixia Liu, Lihui Su, and Shikai Zhang "Chinese electronic medical record named entity recognition model based on pre-training and multi-task learning", Proc. SPIE 12257, 4th International Conference on Information Science, Electrical, and Automation Engineering (ISEAE 2022), 122570R (1 August 2022); https://doi.org/10.1117/12.2640089

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available