Minimum classification error training of landmark models for real-time continuous speech recognition.

Erik McDermott, Timothy J. Hazen

Browse the full ICASSP paper archive.