KenMeSH: Knowledge-enhanced End-to-end Biomedical Text Labelling

TL;DR

KenMeSH is an end-to-end biomedical text labeling model that leverages knowledge-enhanced mask attention to improve large-scale MeSH indexing accuracy.

Contribution

The paper introduces KenMeSH, a novel model that integrates document features with MeSH hierarchy and journal data for improved biomedical indexing.

Findings

Achieves state-of-the-art performance on MeSH indexing tasks.

Effectively incorporates hierarchical and journal information into text labeling.

Outperforms existing methods in accuracy and efficiency.

Abstract

Currently, Medical Subject Headings (MeSH) are manually assigned to every biomedical article published and subsequently recorded in the PubMed database to facilitate retrieving relevant information. With the rapid growth of the PubMed database, large-scale biomedical document indexing becomes increasingly important. MeSH indexing is a challenging task for machine learning, as it needs to assign multiple labels to each article from an extremely large hierachically organized collection. To address this challenge, we propose KenMeSH, an end-to-end model that combines new text features and a dynamic \textbf{K}nowledge-\textbf{en}hanced mask attention that integrates document features with MeSH label hierarchy and journal correlation features to index MeSH terms. Experimental results show the proposed method achieves state-of-the-art performance on a number of measures.