Generalized fast on-the-fly composition algorithm for WFST-based speech recognition.

Takaaki Hori, Atsushi Nakamura

Browse the full Interspeech paper archive.