Inference of variable-length acoustic units for continuous speech recognition.

Sabine Deligne, Frdric Bimbot

Browse the full ICASSP paper archive.