User-adaptive Tourist Information Dialogue System with Yes/No Classifier and Sentiment Estimator

Abstract

We introduce our system developed for Dialogue Robot Competition 2023 (DRC2023). First, rule-based utterance selection and utterance generation using a large language model (LLM) are combined. We ensure the quality of system utterances while also being able to respond to unexpected user utterances. Second, dialogue flow is controlled by considering the results of the BERT-based yes/no classifier and sentiment estimator. These allow the system to adapt state transitions and sightseeing plans to the user.