IBM Watson is now fluent in nine languages

With IBM Watson, it’s possible for computer systems to understand the spoken language. Watson has learned nine languages up to now, which represents more than 90 percent of the world’s business languages. But how does it work that a computer can not just read languages but also understand them and translate them in the right context?

Speech is a natural format for communication all around the globe and even written messages have become increasingly conversational in tone. With IBM Watson it is possible for computer systems to understand spoken languages and the way we express ourselves in text.

How Watson works:

Watson does not only identify words and just reads them, it’s about understanding them. So learning another language is more complex than simply loading a new dictionary.

Source: iStock via Evernine

Language experts taught Watson to understand the structure of languages. They used representative sentences to teach Watson how to identify the verbs, nouns, adjectives and other parts of the sentences. It’s a really complex process because every language has its own compositions.

Watson is now able to speak nine languages: English, Japanese, French, Italian, Spanish, German, Brazilian Portuguese, Arabic and Korean. That represents more than 90% of the world’s business communications and two billion people can be reached in their native language.

Picture source: YouTube / Wired UK
Picture source: YouTube / Wired UK

Watson’s Future

As the number of languages spoken by Watson grows and the sophistication of the understanding increases, new possibilities for applications within and across languages will emerge.