UmlsBERT: Clinical Domain Knowledge Augmentation of Contextual Embeddings Using the Unified Medical Language System Metathesaurus

TL;DR

UmlsBERT enhances contextual embeddings for biomedical NLP by integrating UMLS knowledge during pre-training, leading to improved performance on clinical NER and inference tasks.

Contribution

This paper introduces UmlsBERT, a novel model that incorporates UMLS knowledge into contextual embeddings through a unique augmentation strategy.

Findings

Outperforms existing models on NER tasks

Improves clinical natural language inference accuracy

Effectively encodes domain knowledge into embeddings

Abstract

Contextual word embedding models, such as BioBERT and Bio_ClinicalBERT, have achieved state-of-the-art results in biomedical natural language processing tasks by focusing their pre-training process on domain-specific corpora. However, such models do not take into consideration expert domain knowledge. In this work, we introduced UmlsBERT, a contextual embedding model that integrates domain knowledge during the pre-training process via a novel knowledge augmentation strategy. More specifically, the augmentation on UmlsBERT with the Unified Medical Language System (UMLS) Metathesaurus was performed in two ways: i) connecting words that have the same underlying `concept' in UMLS, and ii) leveraging semantic group knowledge in UMLS to create clinically meaningful input embeddings. By applying these two strategies, UmlsBERT can encode clinical domain knowledge into word embeddings and outperform existing domain-specific models on common named-entity recognition (NER) and clinical natural language inference clinical NLP tasks.