The authors propose a new training approach based on maximum model dis
tance (MMD) for HMMs. MMD uses the entire training set to estimate the
parameters of each HMM, while the traditional maximum likelihood (ML)
only uses those data labelled for the model. Experimental results sho
wed that significant error reduction can be achieved through the propo
sed approach. In addition, the relationship between MMD and corrective
training [3] was discussed, and we have proved that the corrective tr
aining is a special case of MMD approach.