1 paper · 1 filter
Ewan Dunbar, Mathieu Bernard, Nicolas Hamilakis +6
We present the Zero Resource Speech Challenge 2021, which asks participants to learn a language model directly from audio, without any text or labels. The challenge is based on the…