Home » Language Technology

Google Translation part II

11 March 2010

rosetta-stone-flickr-ian-muttoo

A recent article from the New York times discusses the way how Google tries to tackle the translation problem. I wrote about the Google babel phone idea a couple of weeks ago but I wanted to pick up the topic nonetheless. As the article points out Google uses sheer computing power to make automatic translations work. The first approach was to explain the computer all grammar rules of a language, combining it with a dictionary. Unfortunately it seems that there are just too many exemptions and people just don’t like to follow rules (what a surprise 😉 ). Google is now taking already translated documents from the EU and United Nations and comparing these with the text that you enter into the automatic translation tool. By using advanced matching algorithms they can then say which translation fits best. Pretty cool idea, only problem is that most of the spoken / slang / vulgar language still miss the translated documents. But hey, you’ve got to start somewhere. And if you still favour the old approach, there’s a great website for learning Spanish grammar, rules etc.

Image credit: Ian Muttoo, Flickr

One Comment »