Universal Real-Time Translator Now A Reality

Thursday, April 26, 2012

AT&T Research is developing a real-time speech-to-speech translation technology that begins translating as soon as it detects speech, without the latency incurred while waiting for an utterance to complete before translating. The increase in speed is achieved by combining the underlying technologies—automatic speech recognition, natural language understanding, machine learning, speech synthesis—into single step, omitting the usual intermediate step of translating source-language phonemes to target-language text.
AT&T is bringing the universal translator to an app store near you. Well, Granted, it isn’t quite universal yet. For now, the AT&T Translator is capable of seven languages. So far the app supports English, Chinese, French, German, Italian, Japanese and Spanish. It seems reasonable to assume that if the app catches on that more languages will eventually be added.

The app was developed at AT&T Research Labs, using both AT&T Watson and AT&T Natural Voices speech technologies. Usage is simple. Speak a phrase into your device and the Translator converts what you’ve said into the language of your choice before broadcasting it to listeners.
There are countless applications of the technology. Business teleconferences could be handled without requiring a third-party translator, tourists could more easily navigate foreign countries and military units operating in the field could actually talk to the locals without having to shepherd a translator.

Videos of the app are embedded below and available at the AT&T Labs Research website.
According to AT&T, the app uses multiple recognition programs to ensure your translation is as accurate as possible. Redundant recognition programs also analyze each new sentence for context, rather than relying on earlier conversations, which helps avoid a single error at the beginning of a conversation from being automatically included for the duration.

According to AT&T,
Accuracy is also dependent on having a large, complete corpus, and for this the AT&T translation technology depends on statistical methods to extract acoustic, lexical, and translation knowledge from traditional sources—large datasets and existing corpora—as well as nontraditional ones, such as data mining of web pages and their different language versions. This allows the system to be continually and automatically enlarged to cover more domains (health, hotel, entertainment, among others) and keep current with new words and expressions.
The system encompasses multiple translation methods, during the translation, automatically choosing the best one for the two languages being translated. Therefore the system can take advantage of methods that work very fast to convert between two related languages, such as English and Spanish, as well as methods specifically designed for translating between languages that are syntactically dis-similar, such as English and Japanese.


By 33rd SquareSubscribe to 33rd Square