A review of dialogue systems: current trends and future directions

Advances in dialogue systems have recently been made in various fields as an easy to use and inexpensive option to support or replace workers. However, developing dialogue systems that produce satisfactory responses to user queries on par with human workers still presents significant challenges. The primary purpose of this review is to analyse prominent studies on dialogue systems in the literature. Comparison frameworks were developed to perform an in-depth analysis in terms of approaches, data sets and evaluation metrics. Unlike previous reviews, we thoroughly examined how reinforcement learning is applied to dialogue systems. We also analysed studies attempting to interleave the two main types of dialogue systems (i.e. open-domain dialogue and task-oriented dialogue). We present some open-source platforms for developing dialogue systems. Finally, we identified research gaps and discussed potential research directions.