Bangla Question Answering Based Model Using Sequence to Sequence Learning

Abstract

People can use our technology to make their lives easier. In order to keep pace with the current world, we need to automate everything. Sometime it is very difficult to get an immediate answer after asking any question. In this age of technological revolution, everything has become machine dependent. So we can't imagine doing anything without machines now. At present, the models of Deep Learning are working well for Question answering system. This is why we have done our work using deep based Sequence to Sequence model for Bengali context based QA system. Where context and question have been used as part of encoder and answers have been used as a part of decoders. The Question Answering System for NLP is becoming a potentially huge field of research. The main focus of our work is to get answer from question automatically. Question Answering system produce question to answer automatically. Like other languages, we need to expand our research on Bengali language as there is a dearth of materials needed to research Bengali language. We tried to work an automatic question answering system for the Bengali language using general knowledge dataset. For this project, running with the Bengali language turned into very hard task. However end of the day, we've create a model for computerized Question Answering System for the Bangla language.