DTranNER: biomedical named entity recognition with deep learning-based label-label transition model

摘要

Biomedical named-entity recognition (BioNER) automatically identifies specific mentions of interest such as chemicals, diseases, drugs, genes, DNAs, proteins, viruses etc. in biomedical literature. As the fundamental step for various downstream linguistic tasks, e.g., adverse drug event extraction [ ], bacteria biotope task [ ], drug-drug interaction [ ], and protein-protein interaction detection [ ], the performance of BioNER is crucial in the overall biomedical knowledge discovery process [ ].