April 20, 2012
AT&T to Unveil Watson Voice Recognition API for Developers
By Jacqueline Lee
If you could talk to your mobile phone or television and make it obey your requests, would you? With Watson technology from AT&T (News - Alert), that idea will soon become a reality.
AT&T has announced that it will release its Watson voice recognition API to developers sometime in June 2012. Watson, which is real-time speech-to-text software, could feasibly be used to power devices like smart TVs and smartphones as well as powering a wide variety of applications.
“Watson has been used within AT&T for IVR customers, including AT&T's VoiceTone service, for over 20 years during which time the algorithms, tools, and plug-in architecture have been refined to increase accuracy, convenience, and integration,” says AT&T’s Labs Research website. “Besides customer care IVR, AT&T Watson has been used for speech analytics, mobile voice search of multimedia data, video search, voice remote, voice mail to text, web search, and SMS.”
Some iPhone (News - Alert) applications, including Translator, ChaCha and Speak4it already incorporate Watson technology. The software supports HTTP, MRCP and Wireline interfaces. AT&T has worked to perfect Watson so that it recognizes speech regardless of accent or dialect. The release of the API means that developers will not have to construct their own speech recognition capabilities from the ground up.
Sources say that the speech recognition will be divided into categories including SMS, voice dictation and voicemail to text options. These seven categories are necessary because Watson performs better when it knows what types of words to expect. For example, if the user chooses the category “Animals,” then Watson will be able to get itself into the vocabulary ballpark.
The original categories, however, are just the beginning. AT&T hopes to add more with the aid of developers everywhere. AT&T anticipates categories for web search and question and answer as well as categories that will enable Watson to power AT&T’s U-Verse. Watson APIs for gaming and social media are also in development.
In addition to releasing the Watson API, AT&T plans to release the Speech Kit SDK. Speech Kit captures words and phrases and then sends them to servers for translation.
Edited by Juliana Kenny