An improved consensus-like method for Minimum Bayes Risk decoding and lattice combination

Author

Xu, Haihua ; Povey, Daniel ; Mangu, Lidia ; Zhu, Jie

Author_Institution

Dept. of Electron. Eng., Shanghai Jiao Tong Univ., Shanghai, China

fYear

2010

fDate

14-19 March 2010

Firstpage

4938

Lastpage

4941

Abstract

In this paper we describe a method for Minimum Bayes Risk decoding for speech recognition. This is a technique similar to Consensus a.k.a. Confusion Network Decoding, in which we attempt to find the hypothesis that minimizes the Bayes´ Risk with respect to the word error rate, based on a lattice of alternative outputs. Our method is an E-M like technique which makes approximations which we believe are less severe than the approximations made in Consensus, and our experimental results show an improvement in WER both for lattice rescoring and lattice-based system combination, versus baselines such as Consensus, Confusion Network Combination and ROVER.

Keywords

Bayes methods; decoding; speech recognition; E-M like technique; WER; confusion network decoding; consensus-like method; lattice combination; lattice rescoring; lattice-based system combination; minimum Bayes risk decoding; speech recognition; word error rate; Decoding; Dynamic range; Equations; Error analysis; Lattices; Speech recognition; Confusion Network; Consensus; Lattice Rescoring; Minimum Bayes Risk (MBR); Speech Recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics Speech and Signal Processing (ICASSP), 2010 IEEE International Conference on

Conference_Location

Dallas, TX

ISSN

1520-6149

Print_ISBN

978-1-4244-4295-9

Electronic_ISBN

1520-6149

Type

conf

DOI

10.1109/ICASSP.2010.5495100

Filename

5495100