No language left behind: Meta’s new AI model can translate 200 languages
- Mark Zuckerberg is calling it “one of world’s fastest supercomputers.”
- Meta aims to enable over 25 billion translations across Meta’s apps daily.
- While over 7,000 languages are known, many don’t have sufficient data sets to train AI.
Meta’s new AI model can translate 200 different languages, including many that are not supported by present translation systems, reported Daily Mail.
CEO Mark Zuckerberg is calling it “one of the world’s fastest supercomputers.”
The tech giant aims to leave no language behind and wishes to enable over 25 billion translations across Meta’s apps daily via No Language Left Behind (NLLB).
While more than 7,000 languages are known, many do not have sufficient data sets to train AI.
The low resources languages that the company is aiming for include Balinese, Egyptian Arabic, Pangasinan, Sardinian, Nigerian Fulfulde, and Umbundu.
“The AI modelling techniques we used are helping make high-quality translations for languages spoken by billions of people around the world,” Zuckerberg said in a statement.
Meta claims that the new model has the capacity to translate 55 African languages.
Marta R. Costa-jussa, a research scientist at Meta AI said: “This can impact billions of people by allowing them to communicate in their own native language.”
A user researcher at Meta AI said that this could change the way people lived their lives, did their businesses, and educated themselves.