An Active Listening Dialogue Model Focued on “Open Questions” Using Reinforcement Learning
Nana Matsumoto, Kazuaki Ando · 2024
In recent years, the proportion of elderly people in Japan's total population has reached an all-time high, and the number of dementia patients has been increasing at the same time as the aging of the Japanese population. “Validation” is one of the communication methods to attend to the emotions of dementia patients. However, the nursing care industry faces chronic understaffing, making it difficult for caregivers to dedicate sufficient time to communicate with all patients. In this paper, we focus on “open questions” in validation, and investigate how to construct a listening dialogue model that generates responses with a lot of listening using reinforcement learning. In evaluation experiments, we confirmed the effectiveness of reinforcement learning for the construction of the active listening dialogue model.