2020/04/21 by Andriy Mulyar, Mulyar, Andriy, Bridget T. McInnes +1 · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · #Biomedical Text Mining and Ontologies #Computation and Language (cs.CL) #FOS: Computer and information sciences #Information Retrieval (cs.IR) #Machine Learning (cs.LG) #Natural Language Processing Techniques #Topic Modeling
paper · pdf · doi:10.48550/arxiv.2004.10220
openalex publication_date 2020/04/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Clinical notes contain an abundance of important but not-readily accessible\ninformation about patients. Systems to automatically extract this information\nrely on large amounts of training data for which their exists limited resources\nto create. Furthermore, they are developed dis-jointly; meaning that no\ninformation can be shared amongst task-specific systems. This bottle-neck\nunnecessarily complicates practical application, reduces the performance\ncapabilities of each individual solution and associates the engineering debt of\nmanaging multiple information extraction systems. We address these challenges\nby developing Multitask-Clinical BERT: a single deep learning model that\nsimultaneously performs eight clinical tasks spanning entity extraction, PHI\nidentification, language entailment and similarity by sharing representations\namongst tasks. We find our single system performs competitively with all\nstate-the-art task-specific systems while also benefiting from massive\ncomputational benefits at inference.\n