Dual Causal/Non-Causal Self-Attention for Streaming End-to-End Speech Recognition.

Niko Moritz, Takaaki Hori, Jonathan Le Roux

Browse the full Interspeech paper archive.