Internet and society: “Yandex” has developed service of speech recognition in Russian

The company "Yandex" has officially presented a cloud service of speech recognition SpeechKit Cloud, through which developers can make their products understand the voice of a man, according to a company blog.  

At the heart of SpeechKit Cloud is the technology of speech recognition Yandex SpeechKit, which "Yandex" launched in 2013. The technology is involved in about 400 applications for smart phones and tablets for Android, iOS and Windows Phone, and daily processes millions of voice prompts.

The launch of SpeechKit Cloud allows implement human speech recognition not only in mobile applications, but also a variety of programs, services and devices - from computer games to domestic robots and navigation systems for cars.

Currently SpeechKit Cloud recognizes words and phrases in Russian and Turkish. In the future, the developers intend to add support for the English language. The quality of speech recognition depends on the subject of the request. For example, if a person dictates SMS or a note, it is about 82%, if he says a short inquiry - 85%. "Geographical" questions - name and address of the organization, the speech recognizer understands with accuracy close to the abilities of the individual - 95%.

For the processing of information respond servers of "Yandex". On average, to receive a response to the request takes about one second.

SpeechKit Cloud is already used by several partners of the company.