Review of Deep Reinforcement Learning-Based Recommender Systems

Recommender systems (RS) are an indispensable technology that helps to address the issue of information explosion in the digital age. It can assist the customers by proposing personalised items and providers by increasing traffic to their website. Recommender systems can be applied to variety of business use cases including but not limited to electronic commerce applications and media recommendation (e.g., news, motion pictures, music, and videos). Deep reinforcement learning (DRL)-based RS has recently gained popularity as a research topic. Because of its interactive approach and autonomous learning ability, it frequently outperforms classic recommendation approaches, including deep learning-based methods. The objective of this work is to present a thorough overview of the state of the art for DRL implementations in recommendation systems. This paper starts with the background technologies involved in applying DRL in RS. Then, this paper examines recent advancements in DRL-based RS and discusses open issues. This review serves as an introduction to the topic of DRL in RS and its various facets. It identifies potential areas of research and provides valuable feedback to readers from academic and industry.