A Generalized Dynamic Composition Algorithm of Weighted Finite State Transducers for Large Vocabulary Speech Recognition

Author

Cheng, Osbert ; Dines, John ; Doss, M.M.

Author_Institution

IDIAP Res. Inst., Martigny, Switzerland

Volume

4

fYear

2007

fDate

15-20 April 2007

Abstract

We propose a generalized dynamic composition algorithm of weighted finite state transducers (WFST), which avoids the creation of noncoaccessible paths, performs weight look-ahead and does not impose any constraints to the topology of the WFSTs. Experimental results on Wall Street Journal (WSJ1) 20k-word trigram task show that at 17% WER (moderately-wide beam width), the decoding time of the proposed approach is about 48% and 65% of the other two dynamic composition approaches. In comparison with static composition, at the same level of 17% WER, we observe a reduction of about 60% in memory requirement, with an increase of about 60% in decoding time due to extra overheads for dynamic composition.

Keywords

decoding; speech coding; speech recognition; transducers; decoding; generalized dynamic composition algorithm; large vocabulary speech recognition; weighted finite state transducers; Computer science; Costs; Decoding; Heuristic algorithms; Hidden Markov models; Minimization methods; Speech recognition; Topology; Transducers; Vocabulary; Dynamic Composition; Large Vocabulary Continuous Speech Recognition; Weighted Finite State Transducers;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on

Conference_Location

Honolulu, HI

ISSN

1520-6149

Print_ISBN

1-4244-0727-3

Type

conf

DOI

10.1109/ICASSP.2007.366920

Filename

4218108