AI Meta technology can translate up to 200 languages

AI Meta technology

Meta officially announced the latest AI technology model that can translate 200 languages ​​in real time.

Facebook’s parent company, Meta, has reportedly launched a new AI technology that is said to be able to translate hundreds of languages in real time. This research is expected to improve language translation on their social media applications, namely Facebook and Instagram.

Launching from The Verge (7/7), this Artificial Intelligence model is part of Meta’s ‘universal speech translation’ project. This project is important for the company to grow across its various platforms, such as Facebook, Instagram, to develop domains such as VR and AR.

This new AI technology model can speak 200 different languages, including a number of lesser-spoken languages from across Asia and Africa, such as Lao and Kamba. Through this translation, Meta will not only be able to better understand its users, but also enable the technology to form the basis of application foundations in future projects such as their augmented reality glasses.

Engadget revealed, a project called No Language Left Behind (NLLB) was launched by the company six months ago. This project serves to train AI to translate various languages without having to go through English first.

Furthermore, Meta officially announced the huge success of this NLLB project on Wednesday through a blog post. The AI ​​model, dubbed the NLLB-200, can translate in 200 different languages, including 55 African languages, with results that are similar to native speakers.

Meta said that the performance of the AI ​​model on the FLORES-101 benchmark surpassed the existing latest translation model, which was 44 percent. This AI model also focuses on languages ​​that have not been listed in commercial machine translation tools, namely 70 percent for certain African and Indian dialects.

Meta AI technology researcher Angela Fan told The Verge the team was inspired by the lack of attention paid to low-resource languages ​​in commercial machine translation so far. He revealed, the research was conducted to produce a translation technology that is suitable for everyone.

“Translation doesn’t even work for the language we use, so that’s why we started this project,” Fan said.

In line with this, Meta has decided to make the NLBB-200 open source and provide $200,000 in grants to non-profit organizations. This is done to develop the technology into applications that exist in the real world such as Facebook News Feed and Instagram.

“Imagine visiting your favorite Facebook group and then finding a post on Igbo or Luganda, and being able to understand it in your own language with the click of a button,” said Meta in his blog post.

You May Also Like

About the Author: AKDSEO