When I first joined NTT as an intern, I worked on dereverberation technology for voice recordings. This research centered on removing noise and echoes found in sound recordings made with for example IC recorders. Later, I shifted to automatic speech recognition research, and I am now researching speech recognition technology that allows for highly precise recognition even with noise and reverberation interferences. The theme of my research has moved from signal processing to speech recognition technology, but focusing on the problem of noise and reverberation. This means the knowledge I gained from past research is still applicable, and I feel as though the contents of my research has evolved gradually and naturally.
NTT R&D not only employs a number of renowned researchers, but many more visit our company to meet with them, making for ample opportunities to exchange ideas. Of course I actively participate in academic conferences abroad, joining in discussions there, as well as visiting local universities. In fact I have even planned workshops with fellow speech recognition researchers at several universities in Germany, the U.S., Israel etc.




