Google first in MT tests

All those who believe, and hope, that human translators could never be replaced by machines still don't need to panic. However, according to a recent test run by the US government it is no other than search engine giant Google that might become a competitive rival.

The translation tests were conducted by the National Institute of Science and Technology and consisted of translating 100 articles from Agence France Presse and the Xinhua News Agency. According to ZDNet, "Google scored the highest in Arabic-to-English and Chinese-to-English translations."

The success of Google's translation technology, statistical machine translation (SMT), is largely based on the huge amount of data samples collected by the search engine because the results of machine translations improve as more data is available - and Google already stored billions of (mostly human) translated websites - their number growing daily.

Google was well ahead of the University of Southern California's Information Sciences Institute, which came in second, and IBM.

Source: ZDNet, Corante

NIST 2005 Machine Translation Evaluation Official Results