Extensions to phone-state decision-tree clustering: single tree and tagged clustering

Abstract

The article describes two extensions to the "traditional" decision tree methods for clustering allophone HMM states in large vocabulary continuous speech recognition (LVCSR) systems. The first, single tree clustering, combines all allophone states of all phones into a single tree. This can be used to improve the performance for very small systems. The single tree clustering structure can also be exploited for speaker and channel adaptation and is shown to provide a 30% reduction in the error rate for an LVCSR task under matched channel conditions and a greater reduction under mismatched channel conditions. The second, tagged clustering, is a mechanism for providing additional information to the clustering procedure. The tags are labels for any of a wide variety of factors, such as stress, placed on the triphones. These tags are then accessible to the clustering process. Small improvements in the recognition performance were obtained under certain conditions. Both methods can be combined.