CHENNAI: Google has added eight Indian languages, including Sanskrit, to Google Translate. This is part of a growing number of regional languages that Google’s online multilingual translation service supports.
“Sanskrit has been the most requested language on Google Translate and we’re finally adding it,” Isaac Caswell (senior software engineer at Google Research) told ET in an exclusive interview. “We’re also adding languages from northeast India which are a less well-represented area.”
The announcement was made during the annual Google conference I/O, which began Wednesday night.
These languages are the first to be added by the zero-shot translation machine. A machine learning model can only see monolingual text. It does not have an example and learns to translate.
Although this technology is amazing, it’s not perfect. Caswell stated that the models will be improved to provide the same experience as with a Spanish translation or German translation.
Google Translate now supports 133 languages worldwide, and the addition of the Indian languages is part of a larger update that includes 24 languages.
The languages have been added to the vocabulary of more than 300 million people. Mizo, for example, is spoken in northeast India by approximately 800,000 people, while Lingala is used by over 45 million people throughout Central Africa. Google Translate has also been updated to include the addition of three indigenous languages from the Americas (Quechua Guarani, Aymara and Aymara), and one English dialect (Sierra Leonean Krio).