An application of recurrent neural networks to discriminative keyword spotting

Santiago Fernández; Alex Graves; Jürgen Schmidhuber

doi:10.1007/978-3-540-74695-9_23

An application of recurrent neural networks to discriminative keyword spotting

Santiago Fernández, Alex Graves, Jürgen Schmidhuber

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

200 Scopus citations

Abstract

The goal of keyword spotting is to detect the presence of specific spoken words in unconstrained speech. The majority of keyword spotting systems are based on generative hidden Markov models and lack discriminative capabilities. However, discriminative keyword spotting systems are currently based on frame-level posterior probabilities of sub-word units. This paper presents a discriminative keyword spotting system based on recurrent neural networks only, that uses information from long time spans to estimate word-level posterior probabilities. In a keyword spotting task on a large database of unconstrained speech the system achieved a keyword spotting accuracy of 84.5 %. © Springer-Verlag Berlin Heidelberg 2007.

Original language	English (US)
Title of host publication	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Publisher	Springer Verlag
Pages	220-229
Number of pages	10
ISBN (Print)	9783540746935
DOIs	https://doi.org/10.1007/978-3-540-74695-9_23
State	Published - Jan 1 2007
Externally published	Yes

Bibliographical note

Generated from Scopus record by KAUST IRTS on 2022-09-14

Access to Document

10.1007/978-3-540-74695-9_23

Cite this

@inproceedings{a9466dce0858435898727617cec2d518,

title = "An application of recurrent neural networks to discriminative keyword spotting",

abstract = "The goal of keyword spotting is to detect the presence of specific spoken words in unconstrained speech. The majority of keyword spotting systems are based on generative hidden Markov models and lack discriminative capabilities. However, discriminative keyword spotting systems are currently based on frame-level posterior probabilities of sub-word units. This paper presents a discriminative keyword spotting system based on recurrent neural networks only, that uses information from long time spans to estimate word-level posterior probabilities. In a keyword spotting task on a large database of unconstrained speech the system achieved a keyword spotting accuracy of 84.5 %. {\textcopyright} Springer-Verlag Berlin Heidelberg 2007.",

author = "Santiago Fern{\'a}ndez and Alex Graves and J{\"u}rgen Schmidhuber",

note = "Generated from Scopus record by KAUST IRTS on 2022-09-14",

year = "2007",

month = jan,

day = "1",

doi = "10.1007/978-3-540-74695-9_23",

language = "English (US)",

isbn = "9783540746935",

pages = "220--229",

booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

address = "Germany",

}

An application of recurrent neural networks to discriminative keyword spotting. / Fernández, Santiago; Graves, Alex; Schmidhuber, Jürgen.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Springer Verlag, 2007. p. 220-229.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - An application of recurrent neural networks to discriminative keyword spotting

AU - Fernández, Santiago

AU - Graves, Alex

AU - Schmidhuber, Jürgen

N1 - Generated from Scopus record by KAUST IRTS on 2022-09-14

PY - 2007/1/1

Y1 - 2007/1/1

N2 - The goal of keyword spotting is to detect the presence of specific spoken words in unconstrained speech. The majority of keyword spotting systems are based on generative hidden Markov models and lack discriminative capabilities. However, discriminative keyword spotting systems are currently based on frame-level posterior probabilities of sub-word units. This paper presents a discriminative keyword spotting system based on recurrent neural networks only, that uses information from long time spans to estimate word-level posterior probabilities. In a keyword spotting task on a large database of unconstrained speech the system achieved a keyword spotting accuracy of 84.5 %. © Springer-Verlag Berlin Heidelberg 2007.

AB - The goal of keyword spotting is to detect the presence of specific spoken words in unconstrained speech. The majority of keyword spotting systems are based on generative hidden Markov models and lack discriminative capabilities. However, discriminative keyword spotting systems are currently based on frame-level posterior probabilities of sub-word units. This paper presents a discriminative keyword spotting system based on recurrent neural networks only, that uses information from long time spans to estimate word-level posterior probabilities. In a keyword spotting task on a large database of unconstrained speech the system achieved a keyword spotting accuracy of 84.5 %. © Springer-Verlag Berlin Heidelberg 2007.

UR - http://link.springer.com/10.1007/978-3-540-74695-9_23

UR - http://www.scopus.com/inward/record.url?scp=38149014113&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-74695-9_23

DO - 10.1007/978-3-540-74695-9_23

M3 - Conference contribution

SN - 9783540746935

SP - 220

EP - 229

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

PB - Springer Verlag

ER -

An application of recurrent neural networks to discriminative keyword spotting

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this