OMNILINGUAL ASR

Summary from the TLDR AI newsletter

Omnilingual ASR is a speech recognition system that supports over 1,600 languages, with hundreds never before covered by other ASR technology. New languages can be added with just a few paired examples without specialized expertise or large datasets. Omnilingual ASR is designed for broad accessibility - it is aimed at making speech technology more inclusive and adaptable for communities and researchers worldwide.