Phonetic recognition using hidden Markov models and maximum mutual information training (1988)

by B Merialdo
Venue:in Proc. IEEE Int. Conf. Acoustics, Speech, Signal (ICASSP ’88