Facebook Develops a Universal Translation Engine that Comprehends 128 Languages
Remember how in sci-fi movies they have futuristic translation technologies that allow individuals from other universes to talk in real time without needing any preexisting knowledge of the local language?
It could happen soon, thanks to Facebook's AI research team's newest work on an improved language translation system, which can currently translate up to 128 different languages into English within a single app.
The process, known as XLS-R, is capable of faster voice recognition, translation, and language identification than any other system now available.
According to Facebook AI:
“Trained on more than 436,000 hours of publicly available speech recordings, XLS-R is based on wav2vec 2.0, our approach to self-supervised learning of speech representations. Utilizing speech data from different sources, ranging from parliamentary proceedings to audio books, we’ve expanded to 128 different languages, covering nearly two and a half times more languages than its predecessor.”
Matter of fact, Facebook claims that it put XLS-R to the test against four prominent multilingual speech recognition benchmarks, and found that it surpassed them in most cases.
“Specifically, we tried it on five languages of BABEL, 10 languages of CommonVoice, eight languages of MLS, and the 14 languages of VoxPopuli.”
As seen here, these systems' error rates are still quite high for some languages, but the XLS-R demonstrates a significant gain in accuracy, which Facebook is continuing to enhance as the process evolves.
This might potentially open up a slew of new uses, including cross-border connectivity and commerce, allowing enterprises all around the world to increase their reach.
Facebook also mentions that such technologies may be useful in future metaverse applications, allowing for greater universal connection in these digital, open world communities.
As previously stated, Facebook is still working on the system, and Facebook also points out that there are over 7,000 languages spoken worldwide, so it's still a long way off from being a genuinely universal translator. But it's a step in the right direction, and it'll help Facebook become a more important tool for connecting people.
More information regarding Facebook's XLS-R study may be found here.