On using entropy information to improve posterior probability-based confidence measures

Tzan Hwei Chen*, Berlin Chen, Hsin Min Wang

*Corresponding author for this work

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

4 Citations (Scopus)

Abstract

In this paper, we propose a novel approach that reduces the confidence error rate of traditional posterior probability-based confidence measures in large vocabulary continuous speech recognition systems. The method enhances the discriminability of confidence measures by applying entropy information to the posterior probability-based confidence measures of word hypotheses. The experiments conducted on the Chinese Mandarin broadcast news database MATBN show that entropy-based confidence measures outperform traditional posterior probability-based confidence measures. The relative reductions in the confidence error rate are 14.11% and 9.17% for experiments conducted on field reporter speech and interviewee speech, respectively.

Original languageEnglish
Title of host publicationChinese Spoken Language Processing - 5th International Symposium, ISCSLP 2006, Proceedings
Pages454-463
Number of pages10
DOIs
Publication statusPublished - 2006
Event5th International Symposium on Chinese Spoken Language Processing, ISCSLP 2006 - Singapore, Singapore
Duration: 2006 Dec 132006 Dec 16

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4274 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other5th International Symposium on Chinese Spoken Language Processing, ISCSLP 2006
Country/TerritorySingapore
CitySingapore
Period2006/12/132006/12/16

Keywords

  • Confidence measure
  • Continuous speech recognition
  • Entropy
  • Posterior probability

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'On using entropy information to improve posterior probability-based confidence measures'. Together they form a unique fingerprint.

Cite this