
Can Africa’s thousands of languages reboot AI learning?
Across the globe, natural-language processing has remixed language into vectors and tokens, but breakthroughs in AI have largely been trained on English and a handful of dominant tongues. In Saarbrücken, Germany, a researcher named David Ifeoluwa Adelani led a project that rethinks how machines understand Sub-Saharan languages. Working with Saarland University’s Institute for Computational Linguistics…