SMedBERT: A Knowledge-Enhanced Pre-trained Language Model with Structured Semantics for Medical Text Mining

TL;DR

SMedBERT is a medical pre-trained language model that integrates structured semantic knowledge from medical knowledge graphs to improve understanding and performance in medical text mining tasks.

Contribution

Introduces SMedBERT, a novel medical PLM that incorporates deep structured semantic knowledge and a mention-neighbor hybrid attention mechanism for enhanced medical text understanding.

Findings

Significantly outperforms baselines in Chinese medical tasks

Improves question answering, question matching, and natural language inference

Effective knowledge integration enhances semantic representations

Abstract

Recently, the performance of Pre-trained Language Models (PLMs) has been significantly improved by injecting knowledge facts to enhance their abilities of language understanding. For medical domains, the background knowledge sources are especially useful, due to the massive medical terms and their complicated relations are difficult to understand in text. In this work, we introduce SMedBERT, a medical PLM trained on large-scale medical corpora, incorporating deep structured semantic knowledge from neighbors of linked-entity.In SMedBERT, the mention-neighbor hybrid attention is proposed to learn heterogeneous-entity information, which infuses the semantic representations of entity types into the homogeneous neighboring entity structure. Apart from knowledge integration as external features, we propose to employ the neighbors of linked-entities in the knowledge graph as additional global contexts of text mentions, allowing them to communicate via shared neighbors, thus enrich their semantic representations. Experiments demonstrate that SMedBERT significantly outperforms strong baselines in various knowledge-intensive Chinese medical tasks. It also improves the performance of other tasks such as question answering, question matching and natural language inference.