Training of across-word phoneme models for large vocabulary continuous speech recognition.

Achim Sixtus, Hermann Ney

Browse the full ICASSP paper archive.