We participated in the [International Workshop on Spoken Language Translation (IWSLT2017)] held at Waseda University!

We are pleased to announce our exhibition report from the International Workshop on Spoken Language Translation (IWSLT2017) held at Waseda University on Thursday, December 14 - Friday, December 15, 2017.

We exhibited our "Android Voice Recognition + Translation App".
This is a communication app that instantly recognizes and translates text when spoken into a smartphone, and then reads it out loud using text-to-speech synthesis.

We received a lot of interest and questions from visitors to our booth.
One of the most popular questions was about the speed of speech recognition.
The secret of the speed improvement lies in the algorithm!

In the past, acoustic models and languages were generally created by statistics and human hands, but in recent years, the development of neural networks has made it possible to create models using the DNN (deep learning) method.
As a result, we have succeeded in improving the speed and accuracy of speech recognition.

When I actually used this application, I was really surprised at the speed of speech recognition.
I have high expectations for further improvements in speech recognition and translation accuracy in the future when utilizing neural networks!

Also, in talking with visitors at the conference this time,
I would like to use it as a translation tool during meetings and as a tool for taking minutes.
I would like to have a function that can automatically detect the language from speech without having to select the language.
We heard various issues and needs related to translation and speech recognition, such as "I want to use it as a translation tool during meetings and as a tool for taking minutes.

Kodensha will continue to devote itself to daily activities in order to provide the best possible assistance to meet the various needs of each company.

This is our report on the International Workshop on Spoken Language Translation (IWSLT2017).
International Workshop on Spoken Language Translation (IWSLT2017).
Thank you very much for stopping by our booth.