Lattice segmentation and minimum Bayes risk discriminative training
Vlasios Doumpiotis, Stavros Tsakalidis, William J. Byrne · 2003
Lattice segmentation techniques developed for Minimum Bayes Risk decoding in large vocabulary speech recognition tasks are used to compute the statistics needed for discriminative training algorithms that estimate HMM parameters so as to reduce the overall risk over the training data. New estimation procedures are developed and evaluated for both small and large vocabulary recognition tasks, and additive performance improvements are shown relative to maximum mutual information estimation. These relative gains are explained through a detailed analysis of individual word recognition errors. Key words: Discriminative training, maximum mutual information (MMI) estimation, acoustic modeling, minimum Bayes risk decoding, risk minimization, large vocabulary speech recognition, lattice segmentation. 1