One of Google Inc’s vision for the future is to translate documents instantly in the world’s main languages, with machine logic.
The statistical machine translation in Google Research Lab is being worked upon under the guidance of researcher Franz Och and two team members. The statistical machine translation is salient from the fact that it has abstained from feeding grammatical rules and dictionaries as in its previous programs. The novel approach applied is to feed documents that have already been translated into two languages and then rely on computers to determine patterns for future translations. They used the United Nations and European Union documents to train their machine by feeding about 200 billion words.
According to Franz Och ‘the quality is not perfect, but it is an improvement on previous efforts of machine translation.’ He adds ‘the software will not overtake humans in expert translations.’
Many countries with different have language-defined communities and do not understand people from different parts of the globe. The innovation could provide the necessary paradigm shift into the perspectives of the people.
The software might reduce the number of people learning foreign languages. But an internet browser who has acquired knowledge about a foreign country, its people, its culture, just about anything, when he meets a person of that language, he would be at a loss of words.
As in any new venture, the innovative idea by Google does face a few obstacles. They have not found ways to earn money. Some African languages do not have considerable translated text. Though Google chairman Eric Schmidt also sees broad political consequences of a world with easy translations.
Via: Washington Post
