Deep Learning for Conversational Agent via Context Question Answering Model

Abstract

The usage of a conversational device communicate knowledge on behalf of humans has attracted popularity and has been widely employed. As a consequence, this study seeks to analyze and improve the current business platform by providing the ability to offer immediate feedback by integrating Deep Learning to the conversational agent. Our research focuses on Artificial Intelligence (AI) to include Natural Language Processing (NLP) by offering a paradigm focusing on Deep Learning via Context Question Answering Model. This model enables the conversational agent to make wiser and stronger choices to be able to understand and generate their own answers to unknown queries. The testing will be performed in the quantitative approach using Black-box Testing and User Experience (UX) Testing. The Black-box Testing findings show that the average time taken for Deep Learning to produce a response is approximately two seconds. As far as accuracy is concerned, the conversational agent is capable of attaining the accuracy of 67.65 percent. The accuracy could be improved from time to time by providing more training to the model. Apart from this, the test results from User Experience Testing indicates that 84.3 percent of the respondents are satisfied with the conversational agent, whereas 86 percent of respondents would use this conversational agent again. The experiments provide positive outcomes and evidence of the progress of this research, with tremendous contributions to the area of artificial conversational agent.