Should we Hard-Code the Recurrence Concept or Learn it Instead ? Exploring the Transformer Architecture for Audio-Visual Speech Recognition.

George Sterpu, Christian Saam, Naomi Harte

Browse the full Interspeech paper archive.