For some people, Meta is probably the company that is constantly spying on us and then selling our personal data to advertisers. However, Meta’s latest feat indicates that it may not have been so corrupted by the dark side of the Force and is probably more of a gray Jedi than an utterly evil Sith.
in a new one blog postMeta has announced that the latest AI model used to translate content on Instagram, Facebook and the other business technologies can now translate 200 languages. In addition, according to Meta, the new model, called NLLB-200, improves translation quality by an average of 44%. For some African and Indian languages, this percentage even exceeds 70%.
In addition to the new AI model, Meta has also built a dataset called FLORES-200, which will allow researchers to evaluate the NLLB-200. As the company said, the new dataset will help researchers assess the performance of the AI model in 40,000 different language directions and “confirm that the translations are of high quality.”
However, Meta goes even further than that. It opens up its NLLB-200 model and the FLORES-200 dataset to other developers and shares its AI model training and dataset codes with them. Also, Meta has decided to pocket and reward researchers and non-profit organizations who use NLLB-200 for initiatives that support the UN Sustainable Development Goals, such as sustainability, food security, gender equality and education. The award: Up to $200,000 in grants.
NLLB-200 is part of Meta’s No Language Left Behind (NLLB) system. The goal is to enable more and more people to “read content in their preferred language rather than always needing an intermediate language that often misunderstands sentiment or content.” According to Meta, NLLB can help advance other technologies such as voice assistants.
Today, however, the company’s main focus is the metaverse, NLLB can also be used to “build technologies that work well in a wider range of languages”, giving more people access to “immersive experiences in virtual worlds”.