As part of NVIDIA’s GTC Conference, held annually, I was invited to participate in a panel on data-driven approaches to language diversity. I joined NVIDIA Product Manager, Data Strategy, Caroline di Brito Gottlieb, Principal Researcher at Microsoft Research, Kalika Bali, Bonaventure Dossou – PhD student at McGill University and the Mila Quebec AI institute, as well as Damiàn Blasi, Harvard Data Science Fellow at Harvard University.
It was a stimulating and engaging discussion, covering topics such as the axes of diversity in language, such as accent, lexicon, and challenges introduced into speech technologies through factors such as the recording environment – microphones and background noise. Bonaventure spoke about the need to diversity language technologies, particularly to emerging markets such as Africa, which is host to several hundred indigenous languages.
You can see the presentation below (login required).
https://www.nvidia.com/en-us/on-demand/session/gtcspring23-s51447/
