No English translation needed for the Facebook AI model


Technology giant Facebook has announced the world’s first Artificial Intelligence-based translator that will not need English at all to translate over 100 different languages. For instance, if we need to translate Russian to English language and vice versa, training data that is available in the English language is needed for the translation server to respond to web demands.

But Facebook says that its new machine learning translation model dubbed the M2M-100- needs no mid-stage translation language and can thus translate over 98 languages directly.

The Mark Zuckerberg-led company says that all this is happening due to its ML model that stands to be 10 times best that the English-centric ones as more than 7.5 billion sentences across 100 languages were used to create the humongous data set which will be analyzed by over 15 billion parameters.

“For years, AI researchers are working in building a central model that can understand all languages across different tasks. Our M2M fulfills all of these as it can support all languages, dialects, and modalities to serve billions of people in a better way”, says Facebook.

Meanwhile, some unconfirmed reports say that Facebook is all set to use only AI technology to moderate its user’s social media content. So, all its human content moderators are all set to lose their jobs in coming months-A move similar to that of Microsoft where the tech giant choose to replace all its news reporting human contractors (50 from the US and 27 from the UK) with AI propelled machines by June 30 this year.

Naveen Goud
Naveen Goud is a writer at Cybersecurity Insiders covering topics such as Mergers & Acquisitions, Startups, Cyber Attacks, Cloud Security and Mobile Security

No posts to display