Developing an interactive Question Answering System is a challenging task in natural language processing (NLP) since it is used as a benchmark for evaluating the machine’s ability of natural language understanding. These systems struggle when the question answering task is accomplished in multiple turns by the end-users to find a huge amount of information based on what they have previously learned. To resolve this issue, Robertabased Conversational Question Answering System (RoBERTa-CoQAS) is developed to incorporate the conversational history into neural machine comprehension system. By including relevant stories in CoQA dataset this framework will be functional even for children to extract knowledge-oriented information, based on the input stories. The developed system provides the most relevant answer to all the questions which are based on the stories from the CoQA dataset. Experimentation results revealed that the proposed system performs well compared to other models

.