Meta Creates New AI Tool That Recognizes Over Than 4,000 Languages

Introduced as the Massive Multilingual Speech (MMS) project, the AI tool is built with a goal to protect and preserve languages and is now available to the public

Imagine a tool that can translate whatever you write into over 4000 different languages, sounds unreal right? Well, Meta, the parent company to Facebook and Instagram has announced that its AI tool named as the Massive Multilingual Speech (MMS) project can now recognise over 4000 languages.

According to Meta, the Massive Multilingual Speech (MMS) AI tool is created with a goal to preserve and protect languages, their diversity and foster research.

Data suggests that there are around 573 known extinct languages in the world, some of which were major languages used by massive communities in the ancient world. An access to these languages would have helped us decipher lost knowledge and historical facts.

Earth is home to more than 7,000 languages, however around 2,900 or 41% of these languages are endangered, which means that they might soon be extinct alongside all knowledge and history written in those languages.

With tools such as Meta’s Massive Multilingual Speech (MMS), we can preserve and protect these languages and promote diversity.

According to a blog post by Meta, the “Massively Multilingual Speech (MMS) models expand text-to-speech and speech-to-text technology from around 100 languages to more than 1,100, more than ten times as many as before.”

Meta took a unique approach, in order to train its multilingual AI system, the tech giant trained it through different mediums such as using audio recordings of translated religious texts like the bible, helping the AI to get a better hold on languages.

“Collecting audio data for thousands of languages was our first challenge because the largest existing speech datasets cover 100 languages at most. To overcome this, we turned to religious texts, such as the Bible, that have been translated in many different languages and whose translations have been widely studied for text-based language translation research,” said Meta in its blog titled ‘Preserving the World’s Language Diversity Through AI’.

Read more:

Shaheen Afridi – the 1st Z-star of Zero Lifestyle launches premium Fashion Smart Watches