An 86, 000-Word Recognizer Based on Phonemic Models.

Matthew Lennig, Vishwa Gupta, Patrick Kenny, Paul Mermelstein, Douglas D. O'Shaughnessy

Browse the full NAACL paper archive.