Large vocabulary continuous speech recognition using WFST-based linear classifier for structured data

Shinji Watanabe, Takaaki Hori, Atsushi Nakamura

Research output: Contribution to conferencePaper

12 Citations (Scopus)

Abstract

This paper describes a discriminative approach that further advances the framework for Weighted Finite State Transducer (WFST) based decoding. The approach introduces additional linear models for adjusting the scores of a decoding graph composed of conventional information source models (e.g., hidden Markov models and N-gram models), and reviews the WFST-based decoding process as a linear classifier for structured data (e.g., sequential multiclass data). The difficulty with the approach is that the number of dimensions of the additional linear models becomes very large in proportion to the number of arcs in a WFST, and our previous study only applied it to a small task (TIMIT phoneme recognition). This paper proposes a training method for a large-scale linear classifier employed in WFST-based decoding by using a distributed perceptron algorithm. The experimental results show that the proposed approach was successfully applied to a large vocabulary continuous speech recognition task, and achieved an improvement compared with the performance of the minimum phone error based discriminative training of acoustic models.

Original languageEnglish
Pages346-349
Number of pages4
Publication statusPublished - 2010 Dec 1
Externally publishedYes
Event11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010 - Makuhari, Chiba, Japan
Duration: 2010 Sep 262010 Sep 30

Conference

Conference11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010
CountryJapan
CityMakuhari, Chiba
Period10/9/2610/9/30

Keywords

  • Distributed perceptron
  • Large vocabulary continuous speech recognition
  • Linear classifier
  • Speech recognition
  • Weighted finite state transducer

ASJC Scopus subject areas

  • Language and Linguistics
  • Speech and Hearing

Fingerprint Dive into the research topics of 'Large vocabulary continuous speech recognition using WFST-based linear classifier for structured data'. Together they form a unique fingerprint.

  • Cite this

    Watanabe, S., Hori, T., & Nakamura, A. (2010). Large vocabulary continuous speech recognition using WFST-based linear classifier for structured data. 346-349. Paper presented at 11th Annual Conference of the International Speech Communication Association: Spoken Language Processing for All, INTERSPEECH 2010, Makuhari, Chiba, Japan.