Google wants to save languages ​​threatened with machine learning app

The app registers what you are looking at via the camera of your smartphone. Woolaroo then searches the archive for the matching word in one of the ten languages ​​currently part of the app.

Currently, some three thousand languages ​​have the status ‘endangered’. They are usually languages ​​used by small minority groups, whose language is not less and less passed on from parent to child. Fifty percent of the 7,000 languages ​​on Earth are ‘stable’, the rest are national languages ​​that are widely used.

Woolaroo supports Louisian Creole, Calabrian Greek, Māori, Nawat, Tamazight, Sicilian, Yang Zhuang, Rapa Nui, Yiddish, and Yugambeh languages. Speakers of these languages ​​can contribute to the project by submitting words and audio. Moreover, the project is open source “so that any person or organization can build something for their own language,” writes Google.

The app uses the Google Cloud Vision API to analyze images via machine learning. Woolaroo is part of Google’s Arts & Culture platform.


Related Articles

Back to top button