1 paper · 1 filter
Gabriel Pirlogeanu, Dan Oneata, Horia Cucu +1
How can we learn the mapping between written words and their spoken counterparts in the absence of explicit textual supervision? We present a visually grounded method for building…