Abstract | ||
---|---|---|
There is a growing need for dialogue systems that can interact appropriately with children. In some dialogue systems, the system selects an utterance in a given scenario. Our goal is to use prosody to generate an appropriate dialogue without requiring a user's dialogue selection. In this study, we constructed a learning system for an utterance selection in a scenario called "Introduction". The developed system learns the dialogue selections from the prosody of a the mother's selection of dialogues and a child's responses. In our experiments, we found that by using prosodic information for the state, the system can learn new dialogue selections. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1007/978-3-030-96451-1_5 | ADVANCES IN ARTIFICIAL INTELLIGENCE |
Keywords | DocType | Volume |
Dialogue system, Reinforcement learning, Transfer learning, Unsupervised learning | Conference | 1423 |
ISSN | Citations | PageRank |
2194-5357 | 0 | 0.34 |
References | Authors | |
0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yuki Kubo | 1 | 0 | 0.34 |
Natsuki Oka | 2 | 0 | 0.34 |
Subaru Hanada | 3 | 0 | 0.34 |
Kazuaki Tanaka | 4 | 0 | 0.34 |
Tomomi Takahashi | 5 | 0 | 0.34 |