Automated Question Answering Assistant

Abstract

IT firms have a number of clients from various sectors. These vendors have many questions or queries which are to be answered manually. Also the amount of questions asked are huge in numbers approximately in the club of 500 to 600. This includes reading the document thoroughly and extracting all the related information regarding the question and then representing the data in appropriate format required. Answering all such questions manually in a limited period of time is quite a tedious task. This leads to more time consumption and increases the human labour behind it. The project aims at developing an automated system which would create a deep learning model that will input the questions present in any format and answer them automatically with the help of the algorithm and give the output in the required format, thus simplifying the work of searching the answers in a given extract and finding the least error prone answer to the question thus increasing the accuracy. This will also reduce time required behind studying any document and framing the answers from them. This system will be able to answer all types of questions. The system can be used for generating answer keys for online exams. The system will encourage the research that returns answers directly instead of keyword extraction from the documents with ample number of queries. Even it can be used for open domain searching of information over the internet.