A Survey on Modularization of Chatbot Conversational Systems

Chatbots have attracted more and more attention and become one of the hottest technology topics. Deep learning has shown excellent performance in various fields such as image, speech, natural language processing and dialogue, it has greatly promoted the progress of chatbots, it can use large amounts of data to learn response generation and feature representations. Due to the rapid development of deep learning, hand-written rules and templates were quickly replaced by end-to-end neural networks. Neural networks is a powerful model that can solve generation problems in conversation response. People’s requirements for chatbots have also increased with the continuous improvement of neural network models. In this article, we discuss three main technologies in chatbots to meet people’s requirements, syntax analysis, text matching and sentiment analysis, and outline the latest progress and main models of three technologies in the field of chatbots in recent years.