Google announced on Thursday, June 27, that it is adding support for 110 new languages to Google Translate, making it the largest language addition ever. Behind the scenes, the large language model in PaLM 2 previously supported 133 languages. In 2022, it added just 24 new languages using the Zero-Shot machine learning model for machine translation.
The PaLM 2 language model helps you learn to translate new languages, especially when learning languages related to other languages, such as languages close to Hindi such as Awadhi and Marwadi and similar ones including French, Seychellois Creole and Mauritian Creole.
The new languages are spoken by more than 614 million people, some of which are native languages spoken by more than 100 million people, and Google officials said in a later interview that the new languages add to that by at least 1,000,000 people, and hundreds of millions of people will speak multiple languages.
Cantonese is one of the newly added languages. Which has been requested to be added to Google Translate for a long time, but due to the overlapping of writing with Mandarin, this makes it very difficult to obtain data and train models. Moreover, 25% of the newly added languages come from Africa.