Applications of Named Entity Recognition Using Graph Convolution Network

Named Entity Recognition (NER), one of the essential tasks in Natural Language Processing (NLP) has been worked upon by many researchers in the past using varied approaches. One such approach is using the graph neural convolution networks. The advancement of deep learning in graphs made it possible to exploit the dependencies in text using a Graph Convolution Network (GCN) on the dependency tree. This paper explores the work of authors on NER using GCN and what different roles GCN can play within the complete architecture of the model for the benefit of this problem. The GCN model is implemented for three standard datasets including Twitter data and molecular biology dataset and compares with the state of the art. The paper also compares the GCN and non-GCN models and as established on other datasets, finds that the GCN model gives an improvement in performance, which varying a few other factors, was pulled a little further for the CoNLL2003 dataset. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.