Language Translation using Google’s ‘Babel fish’

Even being an advocate for in-person (human) language translation services, it is still fascinating to learn about the advancement of translation services using tools like Google’s Babel fish. There was an interview conducted recently with one of Google’s Translate research scientist that offered some insight in the progress of this tool.

The scientist provided the following information regarding this translation service tool.

• It currently supports 63 languages which is quite a lot. The content for this comes from users inputting translation terms into the Internet.
• There are some challenges which include some of the very tough languages. It’s difficult for this tool to provide language translation for languages that are very different from English. This is because English is used as an intermediate language. Simply put, the tool does not translate Russian to Japanese. It translates Russian to English and then to Japanese.
• Another challenge is that this tool is dependent on its source data. When wanting to add a new language to their list of languages translation services, they have to first verify that there is sufficient data of the language on the Web.
• The scientist does concede that in the future that 80% of languages will be covered by these machine translations, but the other 20% will not be a substitute for a human language interpreter or translator.
  

The scientist was asked about what type of team he has formed to help develop this language translating service tool. He candidly stated that it was Wikipedia, statistics, and his grandmother that helped influence his decision on launching a set of Indian languages. The scientist is originally from India.