Scaled random segmental models

Jacob Goldberger; David Burshtein

doi:10.1109/ICASSP.1998.675388

Scaled random segmental models

Jacob Goldberger, David Burshtein

School of Electrical Engineering

Tel Aviv University

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

Abstract

We present the concept of a scaled random segmental model, which aims to overcome the modeling problem created by the fact that segment realizations of the same phonetic unit differ in length. In the scaled model the variance of the random mean trajectory is inversely proportional to the segment length. The scaled model enables a Baum-Welch type parameter reestimation, unlike the previously suggested, non-scaled models, that require more complicated iterative estimation procedures. In experiments we have conducted with phoneme classification, it was found that the scaled model shows improved performance compared to the non-scaled model.

Original language	English
Title of host publication	Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998
Pages	809-812
Number of pages	4
DOIs	https://doi.org/10.1109/ICASSP.1998.675388
State	Published - 1998
Event	1998 23rd IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998 - Seattle, WA, United States Duration: 12 May 1998 → 15 May 1998

Publication series

Name	ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume	2
ISSN (Print)	1520-6149

Conference

Conference	1998 23rd IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998
Country/Territory	United States
City	Seattle, WA
Period	12/05/98 → 15/05/98

Access to Document

10.1109/ICASSP.1998.675388

Cite this

@inproceedings{f402be869cd1459a826002c474b2eac9,

title = "Scaled random segmental models",

abstract = "We present the concept of a scaled random segmental model, which aims to overcome the modeling problem created by the fact that segment realizations of the same phonetic unit differ in length. In the scaled model the variance of the random mean trajectory is inversely proportional to the segment length. The scaled model enables a Baum-Welch type parameter reestimation, unlike the previously suggested, non-scaled models, that require more complicated iterative estimation procedures. In experiments we have conducted with phoneme classification, it was found that the scaled model shows improved performance compared to the non-scaled model.",

author = "Jacob Goldberger and David Burshtein",

note = "Funding Information: The Abhishek Upadhyay is thankful University Grant Commission (UGC) for providing fellowship for pursuing this work respectively.; 1998 23rd IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998 ; Conference date: 12-05-1998 Through 15-05-1998",

year = "1998",

doi = "10.1109/ICASSP.1998.675388",

language = "אנגלית",

isbn = "0780344286",

series = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

pages = "809--812",

booktitle = "Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998",

}

Goldberger, J & Burshtein, D 1998, Scaled random segmental models. in Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998., 675388, ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, vol. 2, pp. 809-812, 1998 23rd IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998, Seattle, WA, United States, 12/05/98. https://doi.org/10.1109/ICASSP.1998.675388

Scaled random segmental models. / Goldberger, Jacob; Burshtein, David.
Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998. 1998. p. 809-812 675388 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings; Vol. 2).

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

TY - GEN

T1 - Scaled random segmental models

AU - Goldberger, Jacob

AU - Burshtein, David

N1 - Funding Information: The Abhishek Upadhyay is thankful University Grant Commission (UGC) for providing fellowship for pursuing this work respectively.

PY - 1998

Y1 - 1998

N2 - We present the concept of a scaled random segmental model, which aims to overcome the modeling problem created by the fact that segment realizations of the same phonetic unit differ in length. In the scaled model the variance of the random mean trajectory is inversely proportional to the segment length. The scaled model enables a Baum-Welch type parameter reestimation, unlike the previously suggested, non-scaled models, that require more complicated iterative estimation procedures. In experiments we have conducted with phoneme classification, it was found that the scaled model shows improved performance compared to the non-scaled model.

AB - We present the concept of a scaled random segmental model, which aims to overcome the modeling problem created by the fact that segment realizations of the same phonetic unit differ in length. In the scaled model the variance of the random mean trajectory is inversely proportional to the segment length. The scaled model enables a Baum-Welch type parameter reestimation, unlike the previously suggested, non-scaled models, that require more complicated iterative estimation procedures. In experiments we have conducted with phoneme classification, it was found that the scaled model shows improved performance compared to the non-scaled model.

UR - http://www.scopus.com/inward/record.url?scp=84892158814&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.1998.675388

DO - 10.1109/ICASSP.1998.675388

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:84892158814

SN - 0780344286

SN - 9780780344280

T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

SP - 809

EP - 812

BT - Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998

T2 - 1998 23rd IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1998

Y2 - 12 May 1998 through 15 May 1998

ER -

Scaled random segmental models

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this