Jiggasha

Abstract

Question Answering System is a system that allows us to query in different type of questionaries’ in preferred language and it extracts the exact answer for that question. We tried to work on different important fields and issues of question answering systems. We researched on various established question answering systems and explored their qualities which makes them better in their tasks. Bengali is one of the most popular and commonly used languages in the world. Bengali is the native language of Bangladeshis and 2'nd most popular language in India as it's the native language of the people of West Bengal of India. But it is still in its early stages of research regarding Automated Question Answering system in Bengali language. The type of Question Answering System is conversational Machine Learning and it generates answers which are in natural language to questions raised by users that are humans. A huge progress has been seen in the question answering system and its use in a wide variety of tasks. In recent years, impressive progress has been seen in this sector. Using encoder decoder neural architectures which is trained with big data input helps developing efficient QA system. In this research work, initial steps have been taken by us to bring state-of-the-art Question Answering technology using a BERT model in Bangla Language by designing a Question Answering System which is for basic questions about random subjects.