Google has announced its “largest expansion ever” of Google Translate today, adding 110 new languages to the platform. This expansion was made possible by Google’s Palm 2 large language model, which helped Translate more efficiently learn languages closely related to each other. Some of the new additions cover more than 614 million speakers, which will benefit around 8% of the world’s population. A quarter of the new languages added are from Africa, making it the largest expansion of African languages to date.
The new languages added range from major world languages with over 100 million speakers to smaller languages spoken by indigenous communities. For example, Far is a tonal language spoken in Djibouti, Eritrea, and Ethiopia, and had the most voluntary community contributions among the new languages. Cantonese, one of the most requested languages for Google Translate, has also been included in this expansion, despite the difficulty in finding data and training models due to its overlap with Mandarin in writing. Manx, the Celtic language of the Isle of Man, has seen a revival thanks to an island-wide movement, with thousands of speakers now. NKo, a standardized form of the Mandinka languages of West Africa, has a unique alphabet and an active research community developing resources and technology for it.
Other languages added include Punjabi (Shahmukhi) in Perso-Arabic script, tamazight (Amazigh) a Berber language spoken throughout North Africa, and tok pisin, a creole of English origin and the lingua franca of Papua New Guinea. Google aims to support even more language varieties and spelling conventions in the future and build AI models that support the 1,000 most widely spoken languages around the world.
Article Source
https://9to5google.com/2024/06/27/google-translate-new-languages-2/