Matrix parser and its application to HMM-based speech recognition

The authors describe a unified framework for continuous speech recognition (CSR) under grammatical constraints, where trellis calculations and parsing are performed by the same simple fundamental operations, namely multiplication and addition of likelihood matrices. The matrix parser is shown to be a generalization of the Cocke-Younger-Kasami (CYK) parser, which because of its simplicity lends itself to efficient hardware implementation. It also facilitates explicit suprasegmental duration control for all grammatical categories. Preliminary results showed that improved duration control on the mora level raised the recognition accuracy from 86.6% to 88.2%.<>