15. Speech Retrieval

  1. Gokhan Tur5 and
  2. Renato De Mori6
  1. Ciprian Chelba1,
  2. Timothy J. Hazen2,
  3. Bhuvana Ramabhadran3 and
  4. Murat Saraçlar4

Published Online: 23 MAR 2011

DOI: 10.1002/9781119992691.ch15

Spoken Language Understanding: Systems for Extracting Semantic Information from Speech

Spoken Language Understanding: Systems for Extracting Semantic Information from Speech

How to Cite

Chelba, C., Hazen, T. J., Ramabhadran, B. and Saraçlar, M. (2011) Speech Retrieval, in Spoken Language Understanding: Systems for Extracting Semantic Information from Speech (eds G. Tur and R. De Mori), John Wiley & Sons, Ltd, Chichester, UK. doi: 10.1002/9781119992691.ch15

Editor Information

  1. 5

    Microsoft Speech Labs, Microsoft Research, USA

  2. 6

    McGill University, Montreal, Canada and University of Avignon, France

Author Information

  1. 1

    Google, USA

  2. 2

    MIT Lincoln Laboratory, USA

  3. 3

    IBM TJ Watson Research Center, USA

  4. 4

    Boğaziçi University, Turkey

  1. Parts of this chapter have been previously published in Chelba et al. (2008) [© 2008 IEEE]. The authors thank IEEE for granting permission to reproduce some paragraphs, tables and figures in this chapter.

Publication History

  1. Published Online: 23 MAR 2011
  2. Published Print: 25 MAR 2011

ISBN Information

Print ISBN: 9780470688243

Online ISBN: 9781119992691

SEARCH

Keywords:

  • data sets;
  • speech recognition technology;
  • speech retrieval;
  • spoken document retrieval (SDR) system;
  • text document retrieval algorithms

Summary

This chapter discusses the retrieval and browsing of spoken audio documents. It focuses on the application of document search where a user provides a query and the system returns a set of audio documents that best match the query. The primary technical challenges of speech retrieval lie in the retrieval system’s ability to deal with imperfect speech recognition technology that produces errorful output due to misrecognitions caused by inadequate statistical models or out-of-vocabulary words. The chapter provides an overview of the common tasks and data sets, evaluation metrics, and algorithms most commonly used in this growing area of research. In order to properly evaluate the effectiveness of a spoken document retrieval (SDR) system, it is important to understand the nature of the problem that a system is being asked to solve. The chapter provides a brief presentation of state-of-the-art text document retrieval algorithms.

Controlled Vocabulary Terms

document handling; speech recognition