Uttar-DATA: A Neural Bangla Question Answering System

Date
2020
Research Supervisor
Editor
Journal Title
Volume
Issue
Journal Title
Journal ISSN
Volume Title
Abstract
Recent advances in the field of natural language processing have improved state-of-the-art performances on many tasks including extractive question answering or reading comprehension for languages like English. But for low resource languages like Bengali due to lack of data and research on QA similar progress has not been achieved. In this work we use state-of-the-art transformer models to train a fully end to end QA system on a reading comprehension dataset translated from SQuAD 2.0 and compare baseline results of zero shot transfer with fine-tuned transformer models. We also evaluate our models on a human annotated QA validation dataset collected from Bengali Wikipedia.
Description
Keywords
Citation
Department Name
Electrical and Computer Engineering
Publisher
North South University
Printed Thesis
DOI
ISSN
ISBN