- Meta develops AI translation system for a primarily spoken language
- The new AI translator can convert spoken languages like Hokkien into spoken English, helping break down language barriers
- While the model is still in development, it allows someone who speaks Hokkien to communicate with someone who speaks English.
In an effort to break down language barriers, Meta has created a new AI translator that can convert spoken languages such as Hokkien into spoken English. Hokkien, a dialect of southern Min Chinese, is primarily spoken and lacks a standard writing system, making it difficult to develop translation tools for it.
The open-source translation system, which is part of Meta’s Universal Speech Translator (UST) project, has made significant progress in this challenge. The company, formerly known as Facebook, hopes that this, along with other AI methods in development, will eventually allow for real-time speech-to-speech translation across hundreds of languages, including spoken languages.
Languages such as Hokkien are difficult to translate because machine translation tools need a large amount of written text to train on, and such languages lack a widely used writing system. To address this issue, Meta used Mandarin – another Chinese language with a large amount of readily available training data – as a bridge between English and Hokkien.
The project’s researchers also actively collaborated with native Hokkien speakers to validate the accuracy of the AI translation models. Because a speech-to-text system was not available, Meta built the translation system primarily using speech-to-speech translation.
While the model is still in development, it allows someone who speaks Hokkien to communicate with someone who speaks English. The only catch is that in its current state, it can only translate one full sentence at a time. Meta is encouraging other developers to join the party by releasing technologies such as SpeechMatrix to help them create their speech-to-speech translation systems or to build on its work. The company has also made available its Hokkien translation models, as well as the research papers that go with them.
