We are working on improving the performance of speech recognition by improving the model of human voice (acoustic model) using HMM and DNN. We are also working on improving statistical language models.
Spoken dialog interface
One of the challenges of spoken dialogue systems is to prevent the user from feeling unnatural. Therefore, we build a spoken dialogue system that takes into account the timing of the other party and the pitch of the voice.
On the other hand, we also consider the semantic content of the dialogue. In this way, we are building a dialogue system that is robust and responds naturally.
When interacting with speech, people often use pointing and eye contact to convey information. We are trying to realize such human-to-human interaction between humans and machines.
The Kitaoka Lab at Toyohashi University of Technology is always accepting applications for postdoctoral researchers, master’s researchers, other researchers, and technical staff!->Click here for details.
1-1 Unjirigaoka, Tenpaku-cho, Toyohashi, Aichi 441-8580, Japan
Toyohashi University of Technology, Department of Information and Intelligence Engineering
Room F303, Research Building F (Kitaoka)
Tel : 0532-44-6878