Facebook is working on a universal translation engine that can understand 128 different languages
You know those futuristic translation devices that they have in sci-fi movies, which allow characters from totally different universes to converse, in real time, without needing any prior knowledge of the local language?
That may soon be a reality, with Facebook’s AI research team today unveiling their latest work on an improved linguistic translation system, which today can translate up to 128 different languages into English in one app.
Called XLS-R, the process is capable of performing speech recognition, speech translation and language identification at a higher rate than any other similar system available.
As explained by Facebook AI:
“Formed on more than 436,000 hours of publicly available voice recordings, XLS-R is based on wav2vec 2.0, our self-supervised approach to learning speech representations. Using voice data from different sources, ranging from parliamentary debates to audiobooks, we have expanded to 128 different languages, covering almost two and a half times as many languages as its predecessor.
Indeed, Facebook says it tested XLS-R against four main multilingual speech recognition benchmarks, in which it outperformed their results in most of the languages tested.
“Concretely, we tried it on five languages of BABEL, 10 languages of CommonVoice, eight languages of MLS and the 14 languages of VoxPopuli.“
As you can see here, the error rates of these systems are still relatively high for some languages, but the XLS-R shows a marked improvement in accuracy, which Facebook continues to improve as it goes. develop the process.
This could potentially open up a range of new applications, including cross-border connection and commerce, expanding opportunities for businesses around the world.
Facebook also notes that such systems could be of great value in its future metaverse applications, facilitating a more universal connection in these open-world digital communities.
As noted, Facebook is still developing the system, and Facebook also notes that there are over 7,000 languages spoken in the world, so it’s far from a truly universal translator. But that’s another step toward that next step, and toward making Facebook a more login critical utility.
You can read more about Facebook’s XLS-R search here.