UWEE Tech Report Series

Porting Decipher from English to Mandarin


UWEETR-2006-0013

Author(s):
M. Hwang, X. Lei, T. Ng, M. Ostendorf, A. Stolcke, W. Wang, J. Zheng and V. Gadde

Keywords:
speech recognition, Mandarin

Abstract

This paper describes our efforts in porting the SRI Decipher English system into Mandarin for transcribing telephone conversations. This includes all aspects of the system: the pronunciation phone set and lexicon, word segmentation, pitch features, discriminatively trained acoustic models with parameter sharing determined by decision trees, and web-data augmented language models.

Download the PDF version

Download the Gzipped Postscript version