Support vector machines for segmental minimum Bayes risk decoding of continuous speech

Veera Venkataramani; Shantanu Chakrabartty; William Byrne

doi:10.1109/ASRU.2003.1318396

Support vector machines for segmental minimum Bayes risk decoding of continuous speech

Veera Venkataramani, Shantanu Chakrabartty, William Byrne

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

19 Scopus citations

Abstract

Segmental Minimum Bayes Risk (SMBR) Decoding involves the refinement of the search space into sequences of small sets of confusable words. We describe the application of Support Vector Machines (SVMs) as discriminative models for the refined search spaces. We show that SVMs, which in their basic formulation are binary classifiers of fixed dimensional observations, can be used for continuous speech recognition. We also study the use of GiniSVMs, which is a variant of the basic SVM. On a small vocabulary task, we show this two pass scheme outperforms MMI trained HMMs. Using system combination we also obtain further improvements over discriminatively trained HMMs.

Original language	English
Title of host publication	2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	13-18
Number of pages	6
ISBN (Electronic)	0780379802, 9780780379800
DOIs	https://doi.org/10.1109/ASRU.2003.1318396
State	Published - 2003
Event	IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003 - St. Thomas, United States Duration: Nov 30 2003 → Dec 4 2003

Publication series

Name	2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003

Conference

Conference	IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003
Country/Territory	United States
City	St. Thomas
Period	11/30/03 → 12/4/03

Access to Document

10.1109/ASRU.2003.1318396

Cite this

Venkataramani, V., Chakrabartty, S., & Byrne, W. (2003). Support vector machines for segmental minimum Bayes risk decoding of continuous speech. In 2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003 (pp. 13-18). Article 1318396 (2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ASRU.2003.1318396

Venkataramani, Veera ; Chakrabartty, Shantanu ; Byrne, William. / Support vector machines for segmental minimum Bayes risk decoding of continuous speech. 2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003. Institute of Electrical and Electronics Engineers Inc., 2003. pp. 13-18 (2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003).

@inproceedings{4d8d5caeb5c84754b117aae1e35a8dce,

title = "Support vector machines for segmental minimum Bayes risk decoding of continuous speech",

abstract = "Segmental Minimum Bayes Risk (SMBR) Decoding involves the refinement of the search space into sequences of small sets of confusable words. We describe the application of Support Vector Machines (SVMs) as discriminative models for the refined search spaces. We show that SVMs, which in their basic formulation are binary classifiers of fixed dimensional observations, can be used for continuous speech recognition. We also study the use of GiniSVMs, which is a variant of the basic SVM. On a small vocabulary task, we show this two pass scheme outperforms MMI trained HMMs. Using system combination we also obtain further improvements over discriminatively trained HMMs.",

author = "Veera Venkataramani and Shantanu Chakrabartty and William Byrne",

note = "Publisher Copyright: {\textcopyright} 2003 IEEE.; IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003 ; Conference date: 30-11-2003 Through 04-12-2003",

year = "2003",

doi = "10.1109/ASRU.2003.1318396",

language = "English",

series = "2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "13--18",

booktitle = "2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003",

}

Venkataramani, V, Chakrabartty, S & Byrne, W 2003, Support vector machines for segmental minimum Bayes risk decoding of continuous speech. in 2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003., 1318396, 2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003, Institute of Electrical and Electronics Engineers Inc., pp. 13-18, IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003, St. Thomas, United States, 11/30/03. https://doi.org/10.1109/ASRU.2003.1318396

Support vector machines for segmental minimum Bayes risk decoding of continuous speech. / Venkataramani, Veera; Chakrabartty, Shantanu; Byrne, William.
2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003. Institute of Electrical and Electronics Engineers Inc., 2003. p. 13-18 1318396 (2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003).

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

TY - GEN

T1 - Support vector machines for segmental minimum Bayes risk decoding of continuous speech

AU - Venkataramani, Veera

AU - Chakrabartty, Shantanu

AU - Byrne, William

PY - 2003

Y1 - 2003

N2 - Segmental Minimum Bayes Risk (SMBR) Decoding involves the refinement of the search space into sequences of small sets of confusable words. We describe the application of Support Vector Machines (SVMs) as discriminative models for the refined search spaces. We show that SVMs, which in their basic formulation are binary classifiers of fixed dimensional observations, can be used for continuous speech recognition. We also study the use of GiniSVMs, which is a variant of the basic SVM. On a small vocabulary task, we show this two pass scheme outperforms MMI trained HMMs. Using system combination we also obtain further improvements over discriminatively trained HMMs.

AB - Segmental Minimum Bayes Risk (SMBR) Decoding involves the refinement of the search space into sequences of small sets of confusable words. We describe the application of Support Vector Machines (SVMs) as discriminative models for the refined search spaces. We show that SVMs, which in their basic formulation are binary classifiers of fixed dimensional observations, can be used for continuous speech recognition. We also study the use of GiniSVMs, which is a variant of the basic SVM. On a small vocabulary task, we show this two pass scheme outperforms MMI trained HMMs. Using system combination we also obtain further improvements over discriminatively trained HMMs.

UR - http://www.scopus.com/inward/record.url?scp=33645775754&partnerID=8YFLogxK

U2 - 10.1109/ASRU.2003.1318396

DO - 10.1109/ASRU.2003.1318396

M3 - Conference contribution

AN - SCOPUS:33645775754

T3 - 2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003

SP - 13

EP - 18

BT - 2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003

Y2 - 30 November 2003 through 4 December 2003

ER -

Venkataramani V, Chakrabartty S, Byrne W. Support vector machines for segmental minimum Bayes risk decoding of continuous speech. In 2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003. Institute of Electrical and Electronics Engineers Inc. 2003. p. 13-18. 1318396. (2003 IEEE Workshop on Automatic Speech Recognition and Understanding, ASRU 2003). doi: 10.1109/ASRU.2003.1318396

Support vector machines for segmental minimum Bayes risk decoding of continuous speech

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this