Abstract
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.
Original language | English |
---|---|
Pages (from-to) | 6325-6351 |
Number of pages | 27 |
Journal | Neural Computing and Applications |
Volume | 36 |
Issue number | 12 |
DOIs | |
State | Published - Apr 2024 |
Bibliographical note
Publisher Copyright:© The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2023. corrected publication 2024.
Keywords
- Chatbots
- Conversational agents
- Dialogue systems
- Reinforcement learning
ASJC Scopus subject areas
- Software
- Artificial Intelligence