Please wait a minute...
Front. Inform. Technol. Electron. Eng.  2011, Vol. 12 Issue (3): 204-212    DOI: 10.1631/jzus.C1000045
    
An iterative approach to Bayes risk decoding and system combination
Hai-hua Xu, Jie Zhu
Department of Electronic Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
Download:   PDF(309KB)
Export: BibTeX | EndNote (RIS)      

Abstract  We describe a novel approach to Bayes risk (BR) decoding for speech recognition, in which we attempt to find the hypothesis that minimizes an estimate of the BR with regard to the minimum word error (MWE) metric. To achieve this, we propose improved forward and backward algorithms on the lattices and the whole procedure is optimized recursively. The remarkable characteristics of the proposed approach are that the optimization procedure is expectation-maximization (EM) like and the formation of the updated result is similar to that obtained with the confusion network (CN) decoding method. Experimental results indicated that the proposed method leads to an error reduction for both lattice rescoring and lattice-based system combinations, compared with CN decoding, confusion network combination (CNC), and ROVER methods.

Key wordsBayes risk (BR)      Confusion network      Speech recognition      Lattice rescoring      System combination     
Received: 03 March 2010      Published: 09 March 2011
CLC:  TP391.42  
Cite this article:

Hai-hua Xu, Jie Zhu. An iterative approach to Bayes risk decoding and system combination. Front. Inform. Technol. Electron. Eng., 2011, 12(3): 204-212.

URL:

http://www.zjujournals.com/xueshu/fitee/10.1631/jzus.C1000045     OR     http://www.zjujournals.com/xueshu/fitee/Y2011/V12/I3/204


An iterative approach to Bayes risk decoding and system combination

We describe a novel approach to Bayes risk (BR) decoding for speech recognition, in which we attempt to find the hypothesis that minimizes an estimate of the BR with regard to the minimum word error (MWE) metric. To achieve this, we propose improved forward and backward algorithms on the lattices and the whole procedure is optimized recursively. The remarkable characteristics of the proposed approach are that the optimization procedure is expectation-maximization (EM) like and the formation of the updated result is similar to that obtained with the confusion network (CN) decoding method. Experimental results indicated that the proposed method leads to an error reduction for both lattice rescoring and lattice-based system combinations, compared with CN decoding, confusion network combination (CNC), and ROVER methods.

关键词: Bayes risk (BR),  Confusion network,  Speech recognition,  Lattice rescoring,  System combination 
[1] Hong-tao Wang, Yuan-qing Li, Tian-you Yu. Coordinated control of an intelligent wheelchair based on a brain-computer interface and speech recognition[J]. Front. Inform. Technol. Electron. Eng., 2014, 15(10): 832-838.