Abstract
Bidirectional Encoder Representations from Transformers (BERT) and BERT-based approaches are the current state-of-the-art in many natural language processing (NLP) tasks; however, their application to document classification on long clinical texts is limited. In this work, we introduce four methods to scale BERT, which by default can only handle input sequences up to approximately 400 words long, to perform document classification on clinical texts several thousand words long. We compare these methods against two much simpler architectures - a word-level convolutional neural network and a hierarchical self-attention network - and show that BERT often cannot beat these simpler baselines when classifying MIMIC-III discharge summaries and SEER cancer pathology reports. In our analysis, we show that two key components of BERT - pretraining and WordPiece tokenization - may actually be inhibiting BERT's performance on clinical text classification tasks where the input document is several thousand words long and where correctly identifying labels may depend more on identifying a few key words or phrases rather than understanding the contextual meaning of sequences of text.
Original language | English |
---|---|
Article number | 9364676 |
Pages (from-to) | 3596-3607 |
Number of pages | 12 |
Journal | IEEE Journal of Biomedical and Health Informatics |
Volume | 25 |
Issue number | 9 |
DOIs | |
State | Published - Sep 2021 |
Bibliographical note
Publisher Copyright:© 2013 IEEE.
Keywords
- BERT
- clinical text
- deep learning
- natural language processing
- neural networks
- text classification
ASJC Scopus subject areas
- Computer Science Applications
- Health Informatics
- Electrical and Electronic Engineering
- Health Information Management