Efficient Generation of high-order context-dependent Weighted Finite State Transducers for Speech Recognition.

Mike Schuster, Takaaki Hori

Browse the full ICASSP paper archive.