Meta uses AI to translate rare languages
FACEBOOK’S parent company, Meta, is harnessing the power of Artificial Intelligence (AI) to translate some of the world’s rare languages, mainly spoken instead of written, in real-time.
Earlier this year, the company announced plans to teach AI to translate scores of spoken and written languages in real-time. However, Meta recently announced the first speechto-speech translation system powered by AI for Hokkien, a primarily oral language in the Chinese diaspora.
The new translation system forms part of Meta’s Universal Speech Translator project, which has been developing new AI methods to allow real-time speech-to-speech translation across many languages eventually.
The project is set to play a major role in the Metaverse and across Metaowned social networks, including Facebook and Instagram.
However, until the inclusion of the new translation tool, the Universal Speech Translator project focused mainly on written languages, barring the nearly 3 500 living languages that are primarily verbal and lacking a widely used writing system.
Meta said translation tools use large amounts of written text to train an AI model. However, the task is nearly impossible without a written element in languages like Hokkien.
“To address this challenge, we’ve built the first Ai-powered speechto-speech translation system for Hokkien. We’re open-sourcing our Hokkien translation models, evaluation datasets and research papers so others can reproduce and build on our work.”
Meta CEO Mark Zuckerberg said its AI research could help break down language barriers in both the physical and digital worlds, the Metaverse, when it sees greater adoption, with the hopes to encourage connection and mutual understanding.
The announcement comes one year after Facebook Inc rebranded itself to Meta as an indication of its stronger push toward what has become known as the next step in the evolution of social networks – the Metaverse.
The company also has hopes that the Universal Speech Translator project will be used across its social platforms by breaking down language barriers for its other tools like Marketplace, to watch videos in real-time in a preferred language, or to communicate with users who speak different languages.