Nazuki
Appropriate backchannels facilitate dialogue. This is true not only for human-human dialogue, but also for human-agent dialogue. This means that an agent needs to give appropriate backchannels to have a smooth dialogue with a human. We perform multimodal analysis on human-human dialogue corpora and reveal what kind of backchannels are appropriate for the speaker’s facial expression, voice behavior, and utterance content. Using this knowledge, we aim to realize an agent that can give appropriate backchannels.
International conferences
- Naoki Azuma, Daichi Shikama, Asahi Ogushi, Toshiki Onishi, Ryo Ishii, and Akihiro Miyata. What Timing and Behavior Patterns Determine Speed Dating Success? Extended Abstracts of the 2025 ACM CHI Conference on Human Factors in Computing Systems (CHI EA ’25), Article No.xxx, pp.xx–xx (2025).
- Toshiki Onishi, Naoki Azuma, Shunichi Kinoshita, Ryo Ishii, Atsushi Fukayama, Takao Nakamura, and Akihiro Miyata: Prediction of Various Backchannel Utterances Based on Multimodal Information. Proc. the 23rd ACM International Conference on Intelligent Virtual Agents, Article No.47, pp.1–4 (IVA ’23) (2023). [Full-text available]
- Shunichi Kinoshita, Toshiki Onishi, Naoki Azuma, Ryo Ishii, Atsushi Fukayama, Takao Nakamura, and Akihiro Miyata: A Study of Prediction of Listener’s Comprehension Based on Multimodal Information. Proc. the 23rd ACM International Conference on Intelligent Virtual Agents, Article No.30, pp.1–4 (IVA ’23) (2023). [Full-text available]
This project is being conducted with NTT Corp.