A Comparative Review of Expert Systems, Recommender Systems, and Explainable AI
Mudavath Ravi, Atul Negi, Sanjay Chitnis
Abstract
Previously Expert Systems (ES) dominated Artificial Intelligence (AI) applications and various ES were developed in multiple domains. However, due to knowledge acquisition bottlenecks, these systems fell out of use. With the rise in Machine Learning (ML) and Deep Learning (DL) approaches, another category of systems called Recommender Systems (RS) is now developed for various application domains. As ML/DL systems acted like black boxes, explainable AI (XAI) came into the picture to provide explanations for the recommendations or predictions made. In this paper, we review the architectural similarities and differences between these three approaches along with applications and future directions. It is important to study these to predict the future of RS and any possible resurgence of ES, developments in XAI and application domains.