A discriminative model for continuous speech recognition based on Weighted Finite State Transducers.

Shinji Watanabe, Takaaki Hori, Erik McDermott, Atsushi Nakamura

Browse the full ICASSP paper archive.