A weight pushing algorithm for large vocabulary speech recognition

Mehryar Mohri, Michael Riley

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Weighted finite-state transducers provide a general framework for the representation of the components of speech recognition systems; language models, pronunciation dictionaries, contextdependent models, HMM-level acoustic models, and the output word or phone lattices can all be represented by weighted automata and transducers. In general, a representation is not unique and there may be different weighted transducers realizing the same mapping. In particular, even when they have exactly the same topology with the same input and output labels, two equivalent transducers may differ by the way the weights are distributed along each path. We present a weight pushing algorithm that modifies the weights of a given weighted transducer in a way such that the transition probabilities form a stochastic distribution. This results in an equivalent transducer whose weight distribution is more suitable for pruning and speech recognition. We demonstrate substantial improvements of the speed of our recognition system in several tasks based on the use of this algorithm. We report a 45% speedup at 83% word accuracy with a simple single-pass 40; 000-word vocabulary North American Business News (NAB) recognition system on the DARPA Eval '95 test set. With the same technique, we report a 550% speedup at 88% word accuracy in rescoring NAB word lattices with more accurate 2nd-pass models. We finally report a 280% speedup at 68% word accuracy for 100; 000 first name-last name pairs recognition.

Original languageEnglish (US)
Title of host publicationEUROSPEECH 2001 - SCANDINAVIA - 7th European Conference on Speech Communication and Technology
EditorsBorge Lindberg, Henrik Benner, Paul Dalsgaard, Zheng-Hua Tan
PublisherInternational Speech Communication Association
Pages1603-1606
Number of pages4
ISBN (Electronic)8790834100, 9788790834104
StatePublished - 2001
Event7th European Conference on Speech Communication and Technology - Scandinavia, EUROSPEECH 2001 - Aalborg, Denmark
Duration: Sep 3 2001Sep 7 2001

Publication series

NameEUROSPEECH 2001 - SCANDINAVIA - 7th European Conference on Speech Communication and Technology

Other

Other7th European Conference on Speech Communication and Technology - Scandinavia, EUROSPEECH 2001
CountryDenmark
CityAalborg
Period9/3/019/7/01

ASJC Scopus subject areas

  • Communication
  • Linguistics and Language
  • Computer Science Applications
  • Software

Fingerprint Dive into the research topics of 'A weight pushing algorithm for large vocabulary speech recognition'. Together they form a unique fingerprint.

  • Cite this

    Mohri, M., & Riley, M. (2001). A weight pushing algorithm for large vocabulary speech recognition. In B. Lindberg, H. Benner, P. Dalsgaard, & Z-H. Tan (Eds.), EUROSPEECH 2001 - SCANDINAVIA - 7th European Conference on Speech Communication and Technology (pp. 1603-1606). (EUROSPEECH 2001 - SCANDINAVIA - 7th European Conference on Speech Communication and Technology). International Speech Communication Association.