Automatic Linguist (AL) is the name given to Speechmatics’ new machine learning technology that’s capable of learning a whole new language in under one week. So far, the artificial technology (AI) has learned 28 different languages and hopes to one day master all 7,000 from around the world.
The AI was developed in Cambridge using the technology front the University of Cambridge. It works by recognizing patterns in language and then applying this to each new language it encounters. One recent challenge for the company came when asked to produce a system that was able learn Hindi in two weeks. Speechmatics managed to deliver the system, making 23 percent less errors than the leading system out there currently.
“Hindi was surprisingly easy,” said CEO at Speechmatics Benedikt von Thungen. “We discovered it is very similar to English in pronunciation, so we were able to use a process called adaptation. It learned from the different data sets we have.” Each language has different aspects to it that make it unique. Languages like Korean, German, and Finnish all use agglutination where words are joined to form new words. That was just one challenge the company had to solve.
“You have tonal languages, like Vietnamese and Mandarin and its variations, which was another fun challenge. But, it’s a matter of teaching the system to deal with it that opens up a new swathe of languages,” says Benedikt. In less than one day Speechmatics can learn the foundations of a language from a small amount of data through the recognising of grammatical structures and sounds.
The AI platform can be used in a number of different situations. One major use is in providing real-time closed captioning for TV. It could also be used to transcribe meetings and recordings as well as live call recording. “Voice will very quickly become the main mechanism to interact with devices. We are seeing the really early adoption of that with Amazon Alexa, Siri and Google Voice, which is phenomenal but it’s still very one-dimensional,” said Benedikt.
However, the company will soon move into a more intuitive world with more conversational interfaces he confirms. “Our USPs are accuracy, operational performance – memory footprint and speed – the breadth of language coverage and speed with which we offer a new language, and the speed at which we can add new features.”
Source Cambridge Independent