Abstract

The increase in research on conversational systems and their applications, constitutes the main motivation of this work. In this research, a critical analysis of the growth of smart chatbots and their combination with computational intelligence techniques is carried out. The analysis is oriented on three fundamental fronts. First, a review protocol is applied to identify the main schools and centers of knowledge. Then the conversational systems are characterized based on the level of inclusion of artificial intelligence techniques. Finally, the integration of conversational systems with different computational intelligence techniques is reviewed. The analysis identifies that there are many opportunities and lines open to research. In particular, the need to strengthen the application of neutrosophic theory and sets for the evaluation of conversations is identified. Also, the need to combine linguistic data summarization techniques and reinforcement learning is identified to improve training methods and reduce the computational cost of conversational systems responses.