NISHIMURA, Ryota
| Affiliation | Department of Computer Science and Engineering |
|---|---|
| Title | Associate Professor |
| Fields of Research | Spoken Dialog System |
| Degree | Ph.D. (Engineering) |
| nishimura.ryota.tz@ Please append "tut.jp" to the end of the address above. |
|
| Researcher information URL（researchmap） | Researcher information |
Research
I aim to develop natural dialogue systems that enable humans and systems to enjoy conversations themselves. Rather than conventional simple information retrieval systems, I research next-generation spoken dialogue technologies that integrate response timing, prosodic control, and visual expressions like natural human conversations. In particular, I work on developing systems that detect response timing in real time and determine appropriate response phenomena such as "back-channels," "repetitions," and "information provision." I also promote the social implementation of research results through elderly-oriented speech technology, multilingual support, and practical platform development.
Theme1：Response Timing Control Technology
Overview
We develop systems that analyze natural response timing in human-to-human dialogues and detect response timing in real time. Using decision tree-based response determination models, we judge appropriate response phenomena such as "back-channels," "repetitions," and "information provision" at 100ms intervals, utilizing prosodic information such as pitch and power gradients, utterance length, and pause length.