The CommonLingua program supports 61 African languages
Pleias and GSMA have released an open-source language identification program based on artificial intelligence, CommonLingua, which is designed for large-scale access to data in African languages.
There are more than 2,000 living languages in Africa, many of which are underrepresented in artificial intelligence training data. As a result, language identification systems do not work very reliably with content in African languages, especially when distinguishing closely related texts.
CommonLingua is designed to remedy this situation. The new program covers 61 African languages from eight language families, including Bantu (21), Niger-Congo (18), Semitic (7) and others. CommonLingua provides consistent text processing in all alphabets used in Africa, including Latin, Arabic, and Ethiopian.
TECHAFRICANEWS
