About Our Laboratory
We are working on speech information processing with a focus on speech recognition.
- Numerical Anaysis (E) supplemental material inverseiteration.c
- Uploaded slides for invited talk at SP October meeting. (CSJ recipe for Kaldi)
- Uploaded slides for invited talk at GTC2016 conference. (Evolutionary optimization of ASR systems)
- Tomohiro Tanaka, a master student, received an excellent student presentation award from the Acoustic Society of Japan.
- Uploaded speech recognition lecture slides
- Gave a lecture in a seminar held at Kyoto University
Attended Frederick Jelinex Memorial Summer Workshop
- Updated Kaldi CSJ recipe.
- Uploaded LSTM-LM meta-prameters optimized by evolutionary computation peformed on Tsubame super computer.
- Remodeled the homepage following the education reform of Tokyo Tech.
- Takafumi Moriya, a master student, received an excellent student presentation award from the Acoustic Society of Japan.
- Tutorial materials for the Kaldi CSJ recipe is available (in Japanese).
- Our proposal has been accepted for 2016 Jelinek Summer Workshop in Speech and Language Technology.
- A Kaldi based recipe is released for Japanese large vocabulary spontaneous speech recognition using the Corpus of Spontaneous Japanese (CSJ). The recipe is developed by a cooperation of our lab and Mitsubishi Electric Research Lab in the US. It is included in the official Kaldi package under egs/csj.