Meta’s AI models are able to recognize and produce more than 1,000 different languages
The development of speech applications for more languages could be possible.
Meta has developed AI models capable of recognizing and producing speech in more than 1,000 different languages. This is a tenfold improvement over what is currently available. The company claims that it’s an important step in preserving languages at risk of extinction.
Meta releases its models via the code-hosting service GitHub. The company claims that by making the models open source, it will allow developers to work in different languages and create new speech applications. For example, messaging services that are universally understood or virtual reality systems that can be used anywhere.
Source:
https://www.technologyreview.com/2023/05/22/1073471/metas-new-ai-models-can-recognize-and-produce-speech-for-more-than-1000-languages/
