Building a hospitable and reliable dialogue system for android robots: a scenario-based approach with large language models
Takato Yamazaki,Katsumasa Yoshikawa,Toshiki Kawamoto,Tomoya Mizumoto,Masaya Ohagi,Toshinori Sato
DOI: https://doi.org/10.1080/01691864.2023.2244554
IF: 2.057
2023-08-24
Advanced Robotics
Abstract:Dialogue systems implemented with android robots are expected to provide not only advanced conversational ability but also reliability and hospitality due to their human-like appearance. In this study, we aim to develop a hospitable dialogue system by encouraging open-ended utterances and responding adaptively to give users the feeling of being heard. To achieve this, utilizing large language models (LLMs) is a promising option, but task-oriented dialogue systems implemented with only LLMs often generate irrelevant, inconsistent, or contradictory utterances. Therefore, we propose a scenario-based dialogue system that subdivides the task into smaller sub-tasks, such as summarization, information extraction, and response generation, and uses LLMs in a fine-grained manner to overcome such shortcomings of the LLM-based dialogue system. Our system was evaluated in the tourist-spot recommendation task of the Dialogue Robot Competition 2022 and achieved second place in the preliminary round and first place in the final round, outperforming other rule-based dialogue systems. However, we also identified several challenges when using LLMs for android dialogue systems, including response delays due to computational complexity, hallucinations, and coordination issues between generated utterances and robot control. GRAPHICAL
robotics