Description
Obtaining large datasets to take advantage of highly expressive deep learning methods is difficult in clinical natural language processing. We address this difficulty by pre-training a clinical text encoder on billing code data. We explore several neural encoder architectures and deploy the text representations obtained from these encoders in the context of clinical text classification tasks. While our ultimate goal is learning a universal clinical text encoder, we also experiment with training a phenotype-specific encoder.
Learning Objective: Our objective is to develop algorithms for encoding clinical text into universal representations that can be used for a variety of phenotyping tasks.
Authors:
Dmitriy Dligach (Presenter)
Loyola University Chicago
Majid Afshar, Loyola University School of Medicine
Timothy Miller, Boston Children's Hospital
Presentation Materials: