Relation Extraction using Using Natural Language Processing and Deep Learning Techniques

O'Donohoe, Bryan (2022) Relation Extraction using Using Natural Language Processing and Deep Learning Techniques. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (770kB) | Preview

Abstract

This research study aims to address the issue of relation extraction. The area of relation extraction has various use cases such as question and answering systems, fact checking and the conversion of semi-structured and unstructured text into knowledge bases. There is now more information than ever before available online, and with this comes additional need for machines to translate this data into easily searchable databases. This study describes the approach taken in the end-to-end deployment of a model in which the end user can extract a relation between two named entities on an AWS hosted server. The dataset used to develop this system is taken from the Wikipedia website and contains labelled text for two named entities. The key findings in this report are that a bidirectional encoder representations from transformers model is the optimal solution when trying to extract relations between two named entities from a body of text. There is scope for further research with the development of a hybrid generative and discriminative model, as well as a more optimal deployment of the final application.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Tamara Malone
Date Deposited:	27 Feb 2023 15:22
Last Modified:	02 Mar 2023 08:29
URI:	https://norma.ncirl.ie/id/eprint/6242

Actions (login required)

View Item