What is going on
Meta’s No Language Left Behind AI project can now translate 200 different languages.
Why it matters
For many people around the world who speak rare or less commonly used languages, better translation tools promise to make the internet a more inclusive place for them to communicate and express themselves.
Meta will use No Language Left Behind to improve translation on Facebook and Instagram before finally implementing it in the metaverse. It also makes it open source for anyone to use.
In our increasingly globalized online world, it is essential to have effective translation tools to communicate across borders, cultures and languages. For native speakers of languages such as English and Spanish, these tools are now quite advanced, but the less often a language is spoken, the less likely native speakers of those languages will be able to use these tools.
That’s where Meta’s No Languages Left Behind project comes in to build an AI model that can translate more and more languages, even the ones that are less commonly spoken. First announced in March, the AI model can now translate into 200 languages, the tech giant said Thursday.
He called it an “AI superpower,” said Meta CEO Mark Zuckerberg in a Facebook post that it will not only be used in Meta’s various products and services, but will also be available on github for someone else to use.
“We just developed an AI model that we built that can translate into 200 different languages, many of which are not supported by current translation systems,” he said. “We call this project No Language Left Behind and the AI modeling techniques we’ve used from NLLB help us create high-quality translations on Facebook and Instagram for languages spoken by billions of people around the world.”
Meta invests heavily in AI research, with hubs of scientists around the world buildingand tools to among many others † With this investment, the company can ensure it remains at the forefront of innovation by partnering with the best AI researchers, while also maintaining a connection with the wider research community through open sourcing projects such as No Languages Left Behind. .
The biggest challenge in creating a translation model that works in rarer languages is that the researchers have a much smaller amount of data — in this case sample sentences — to train the model compared to, say, English. In many cases, they had to find people who spoke those languages to help them provide the data and then check that the translations were correct.
No Language Left Behind can be immediately applied to existing social Meta platforms. But company researchers believe it’s in the metaverse —— where the tool really comes into its own. “Billions of people around the world don’t have access to a technology or translation service that actually works well for their language,” Angela Fan, a meta AI research scientist, said in a video created by the company. “We really hope that the technology we develop will make the metaverse inclusive by design.”