Labrador: Exploring the Limits of Masked Language Modeling for Laboratory Data

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bellamy, David R., Kumar, Bhawesh, Wang, Cindy, Beam, Andrew
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915048421588992
author Bellamy, David R.
Kumar, Bhawesh
Wang, Cindy
Beam, Andrew
author_facet Bellamy, David R.
Kumar, Bhawesh
Wang, Cindy
Beam, Andrew
contents In this work we introduce Labrador, a pre-trained Transformer model for laboratory data. Labrador and BERT were pre-trained on a corpus of 100 million lab test results from electronic health records (EHRs) and evaluated on various downstream outcome prediction tasks. Both models demonstrate mastery of the pre-training task but neither consistently outperform XGBoost on downstream supervised tasks. Our ablation studies reveal that transfer learning shows limited effectiveness for BERT and achieves marginal success with Labrador. We explore the reasons for the failure of transfer learning and suggest that the data generating process underlying each patient cannot be characterized sufficiently using labs alone, among other factors. We encourage future work to focus on joint modeling of multiple EHR data categories and to include tree-based baselines in their evaluations.
format Preprint
id arxiv_https___arxiv_org_abs_2312_11502
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Labrador: Exploring the Limits of Masked Language Modeling for Laboratory Data
Bellamy, David R.
Kumar, Bhawesh
Wang, Cindy
Beam, Andrew
Computation and Language
Artificial Intelligence
Machine Learning
In this work we introduce Labrador, a pre-trained Transformer model for laboratory data. Labrador and BERT were pre-trained on a corpus of 100 million lab test results from electronic health records (EHRs) and evaluated on various downstream outcome prediction tasks. Both models demonstrate mastery of the pre-training task but neither consistently outperform XGBoost on downstream supervised tasks. Our ablation studies reveal that transfer learning shows limited effectiveness for BERT and achieves marginal success with Labrador. We explore the reasons for the failure of transfer learning and suggest that the data generating process underlying each patient cannot be characterized sufficiently using labs alone, among other factors. We encourage future work to focus on joint modeling of multiple EHR data categories and to include tree-based baselines in their evaluations.
title Labrador: Exploring the Limits of Masked Language Modeling for Laboratory Data
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2312.11502