Automatic regularization of cross-entropy cost for speaker recognition fusion

Ville Hautamäki; Kong Aik Lee; David Van Leeuwen; Rahim Saeidi; Anthony Larcher; Tomi Kinnunen; Taufiq Hasan; Seyed Omid Sadjadi; Gang Liu; Hynek Boril; John H.L. Hansen; Benoit Fauve

Automatic regularization of cross-entropy cost for speaker recognition fusion

Ville Hautamäki
, Kong Aik Lee
, David Van Leeuwen
, Rahim Saeidi
, Anthony Larcher
, Tomi Kinnunen
, Taufiq Hasan
, Seyed Omid Sadjadi
, Gang Liu
, Hynek Boril
, John H.L. Hansen
, Benoit Fauve

Research output: Journal article publication › Conference article › Academic research › peer-review

Abstract

In this paper we study automatic regularization techniques for the fusion of automatic speaker recognition systems. Parameter regularization could dramatically reduce the fusion training time. In addition, there will not be any need for splitting the development set into different folds for cross- validation. We utilize majorization-minimization approach to automatic ridge regression learning and design a similar way to learn LASSO regularization parameter automatically. By experiments we show improvement in using automatic regularization.

Original language	English
Pages (from-to)	1609-1613
Number of pages	5
Journal	Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
Publication status	Published - Aug 2013
Externally published	Yes
Event	14th Annual Conference of the International Speech Communication Association, INTERSPEECH 2013 - Lyon, France Duration: 25 Aug 2013 → 29 Aug 2013

ASJC Scopus subject areas

Language and Linguistics
Human-Computer Interaction
Signal Processing
Software
Modelling and Simulation

Cite this

@article{6166acbbbbf644ce9644cf682885f733,

title = "Automatic regularization of cross-entropy cost for speaker recognition fusion",

abstract = "In this paper we study automatic regularization techniques for the fusion of automatic speaker recognition systems. Parameter regularization could dramatically reduce the fusion training time. In addition, there will not be any need for splitting the development set into different folds for cross- validation. We utilize majorization-minimization approach to automatic ridge regression learning and design a similar way to learn LASSO regularization parameter automatically. By experiments we show improvement in using automatic regularization.",

author = "Ville Hautam{\"a}ki and Lee, \{Kong Aik\} and \{Van Leeuwen\}, David and Rahim Saeidi and Anthony Larcher and Tomi Kinnunen and Taufiq Hasan and Sadjadi, \{Seyed Omid\} and Gang Liu and Hynek Boril and Hansen, \{John H.L.\} and Benoit Fauve",

year = "2013",

month = aug,

language = "English",

pages = "1609--1613",

journal = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",

issn = "2308-457X",

publisher = "International Speech Communication Association",

note = "14th Annual Conference of the International Speech Communication Association, INTERSPEECH 2013 ; Conference date: 25-08-2013 Through 29-08-2013",

}

Automatic regularization of cross-entropy cost for speaker recognition fusion. / Hautamäki, Ville; Lee, Kong Aik; Van Leeuwen, David et al.
In: Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, 08.2013, p. 1609-1613.

Research output: Journal article publication › Conference article › Academic research › peer-review

TY - JOUR

T1 - Automatic regularization of cross-entropy cost for speaker recognition fusion

AU - Hautamäki, Ville

AU - Lee, Kong Aik

AU - Van Leeuwen, David

AU - Saeidi, Rahim

AU - Larcher, Anthony

AU - Kinnunen, Tomi

AU - Hasan, Taufiq

AU - Sadjadi, Seyed Omid

AU - Liu, Gang

AU - Boril, Hynek

AU - Hansen, John H.L.

AU - Fauve, Benoit

PY - 2013/8

Y1 - 2013/8

N2 - In this paper we study automatic regularization techniques for the fusion of automatic speaker recognition systems. Parameter regularization could dramatically reduce the fusion training time. In addition, there will not be any need for splitting the development set into different folds for cross- validation. We utilize majorization-minimization approach to automatic ridge regression learning and design a similar way to learn LASSO regularization parameter automatically. By experiments we show improvement in using automatic regularization.

AB - In this paper we study automatic regularization techniques for the fusion of automatic speaker recognition systems. Parameter regularization could dramatically reduce the fusion training time. In addition, there will not be any need for splitting the development set into different folds for cross- validation. We utilize majorization-minimization approach to automatic ridge regression learning and design a similar way to learn LASSO regularization parameter automatically. By experiments we show improvement in using automatic regularization.

UR - https://www.scopus.com/pages/publications/84906262523

M3 - Conference article

AN - SCOPUS:84906262523

SN - 2308-457X

SP - 1609

EP - 1613

JO - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

JF - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

T2 - 14th Annual Conference of the International Speech Communication Association, INTERSPEECH 2013

Y2 - 25 August 2013 through 29 August 2013

ER -

Automatic regularization of cross-entropy cost for speaker recognition fusion

Abstract

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this