The Transducer (e.g. RNN-Transducer or Conformer-Transducer) generates a...
Modern speech recognition systems rely on self-attention. Unfortunately,...
Federated learning with differential privacy, i.e. private federated lea...
Federated learning with differential privacy, or private federated learn...
We describe the design of our federated task processing system. Original...
Information on speaker characteristics can be useful as side information...