AT&T is giving developers the tools to build apps and services that use your voice to communicate with smartphones, televisions and other digital devices.
More than a million hours of research and development has already gone into AT&T's pioneering Watson speech recognition technology that has already been powering advanced speech services in the marketplace for decades. For example it powers their mobile directory search and voice mail to text services.
Now AT&T has announced that it plans to make available several of its Watson Speech APIs. Presumably the name relates to Edison's assistant who received the first phone call "Mr. Watson come here. I need you". Also not to be confused with IBM's Watson question answering program.
As well as the API it plans to release a speech kit SDK to capture a user’s spoken words and send them into the network for transcription, for other developers to incorporate voice recognition and transcription capabilities in their own apps.
The first APIs, coming in June, according to the announcement will be focused on seven areas:
local business search
question and answer
voice mail to text
U-verse electronic programming guide,
Later on we can expect APIs for gaming, social media and other areas. The speech recognition technology uses the subject area to improve its accuracy by using an a restricted set of target words.
You can see what AT&T think of the technology in the following promo video:
Recently a paper reported the information that neural networks seem to have a fundamental problem in recognizing things now we have another twist on the same basic idea. In this case you can construct [ ... ]