The Tokyo Institute of Technology WFST Decoder

From HLT@INESC-ID

Diamantino Caseiro

Date

  • 14:30, February 02, 2007
  • 3rd floor meeting room

Speaker

Abstract

In this talk preliminary evaluations will be presented on the large vocabulary speech decoder being developed at Tokyo Institute of Technology. The goal is to build a scalable and flexible decoder to operate on weighted finite-state transducer (WFST) search spaces. Even though the development of the decoder is still in its infancy we are already achieving favorable accuracy and speed on a large vocabulary spontaneous speech recognition task.