Date of Original Version
Abstract or Description
We describe an extension to the Baum-Welch algorithm for training Hidden Markov Models that uses explicit phoneme segmentation to constrain the forward and backward lattice. The HMMs trained with this algorithm can be shown to improve the accuracy of automatic phoneme segmentation. In addition, this algorithm is signiﬁcantly more computationally efﬁcient than the full BaumWelch algorithm, while producing models that achieve equivalent accuracy on a standard phoneme recognition task.