- Meta recently launched a new AI Model named NNLB-200 that will translate 200 languages.
- The new technology is capable of improving the translation by an average of 44%.
- A dataset named FLORES-200 is also built to evaluate and improve NLLB-200 performance.
To overcome the language barrier and help people connect better today and also be part of the metaverse of tomorrow, the Meta team and their AI researchers have created No Language Left Behind (NNLB), an initiative to provide effective machine translation tools for the majority of the world’s languages.
The new model announcement was made on the official Twitter handle of Meta Newsroom. Also, the CEO, Mark Zuckerberg, posted about the new and latest project and its perks and how will the new technology will change millions of lives over his FB account.
The team has developed a single AI model named NLLB-200, which will be capable of translating 200 different languages and whose outcomes are significantly more precise than what was possible with earlier technologies.
Our latest AI model can translate 200 languages with far greater accuracy than previous technology. It will help billions of people connect better online and make virtual experiences more accessible as well.— Meta Newsroom (@MetaNewsroom) July 6, 2022
Mark Zuckerberg just shared more on this 👇https://t.co/xcJjm7BOot pic.twitter.com/uEvcclGWAP
Now, because there are hundreds of languages for which there are no high-quality translation technologies, billions of people today are unable to access digital material or engage fully in online discussions and communities in their preferred or native languages.
The translation quality of NLLB-200 has been, on average, 44% better than that of earlier AI studies. The translations produced by NLLB-200 were more precise for some languages with African and Indian origins.
A dataset, FLORES-200, is also built to evaluate and improve NLLB-200; this will allow researchers to assess the AI model’s performance in 40,000 different language directions.
FLORES-200 will allow the team to calculate NLLB-200’s performance and execution in each language to make sure that the translations are high-quality.