Relevance vector machines with empirical likelihood-ratio kernels for PLDA speaker verification

Wei Rao; Man Wai Mak

doi:10.1109/ISCSLP.2014.6936591

Relevance vector machines with empirical likelihood-ratio kernels for PLDA speaker verification

The Hong Kong Polytechnic University

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

1 Citation (Scopus)

Abstract

Previous works have shown the benefits of empirical likelihood ratio (LR) kernels for i-vector/PLDA speaker verification. The method not only utilizes the multiple enrollment utterances of target speakers effectively, but also opens up opportunity for adopting sparse kernel machines for PLDA-based speaker verification systems. This paper proposes taking the advantages of the empirical LR kernels by incorporating them into relevance vector machines (RVMs). Results on NIST 2012 SRE demonstrate that the performance of RVM regression equipped with empirical LR kernels is slightly better than that of the support vector machines after performing utterance partitioning.

Original language	English
Title of host publication	Proceedings of the 9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014
Publisher	IEEE
Pages	64-68
Number of pages	5
ISBN (Electronic)	9781479942206
DOIs	https://doi.org/10.1109/ISCSLP.2014.6936591
Publication status	Published - 24 Oct 2014
Event	9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014 - Singapore, Singapore Duration: 12 Sept 2014 → 14 Sept 2014

Conference

Conference	9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014
Country/Territory	Singapore
City	Singapore
Period	12/09/14 → 14/09/14

Keywords

Empirical LR kernel
I-vectors
NIST SRE
Probabilistic linear discriminant analysis
Relevance vector machines

ASJC Scopus subject areas

Information Systems
Computer Science Applications
Software

More information

10.1109/ISCSLP.2014.6936591

Cite this

@inproceedings{e11d2b1d7a6240e5bed0092edb633fae,

title = "Relevance vector machines with empirical likelihood-ratio kernels for PLDA speaker verification",

abstract = "Previous works have shown the benefits of empirical likelihood ratio (LR) kernels for i-vector/PLDA speaker verification. The method not only utilizes the multiple enrollment utterances of target speakers effectively, but also opens up opportunity for adopting sparse kernel machines for PLDA-based speaker verification systems. This paper proposes taking the advantages of the empirical LR kernels by incorporating them into relevance vector machines (RVMs). Results on NIST 2012 SRE demonstrate that the performance of RVM regression equipped with empirical LR kernels is slightly better than that of the support vector machines after performing utterance partitioning.",

keywords = "Empirical LR kernel, I-vectors, NIST SRE, Probabilistic linear discriminant analysis, Relevance vector machines",

author = "Wei Rao and Mak, {Man Wai}",

year = "2014",

month = oct,

day = "24",

doi = "10.1109/ISCSLP.2014.6936591",

language = "English",

pages = "64--68",

booktitle = "Proceedings of the 9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014",

publisher = "IEEE",

address = "United States",

note = "9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014 ; Conference date: 12-09-2014 Through 14-09-2014",

}

Rao, W & Mak, MW 2014, Relevance vector machines with empirical likelihood-ratio kernels for PLDA speaker verification. in Proceedings of the 9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014., 6936591, IEEE, pp. 64-68, 9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014, Singapore, Singapore, 12/09/14. https://doi.org/10.1109/ISCSLP.2014.6936591

Relevance vector machines with empirical likelihood-ratio kernels for PLDA speaker verification. / Rao, Wei; Mak, Man Wai.
Proceedings of the 9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014. IEEE, 2014. p. 64-68 6936591.

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - Relevance vector machines with empirical likelihood-ratio kernels for PLDA speaker verification

AU - Rao, Wei

AU - Mak, Man Wai

PY - 2014/10/24

Y1 - 2014/10/24

N2 - Previous works have shown the benefits of empirical likelihood ratio (LR) kernels for i-vector/PLDA speaker verification. The method not only utilizes the multiple enrollment utterances of target speakers effectively, but also opens up opportunity for adopting sparse kernel machines for PLDA-based speaker verification systems. This paper proposes taking the advantages of the empirical LR kernels by incorporating them into relevance vector machines (RVMs). Results on NIST 2012 SRE demonstrate that the performance of RVM regression equipped with empirical LR kernels is slightly better than that of the support vector machines after performing utterance partitioning.

AB - Previous works have shown the benefits of empirical likelihood ratio (LR) kernels for i-vector/PLDA speaker verification. The method not only utilizes the multiple enrollment utterances of target speakers effectively, but also opens up opportunity for adopting sparse kernel machines for PLDA-based speaker verification systems. This paper proposes taking the advantages of the empirical LR kernels by incorporating them into relevance vector machines (RVMs). Results on NIST 2012 SRE demonstrate that the performance of RVM regression equipped with empirical LR kernels is slightly better than that of the support vector machines after performing utterance partitioning.

KW - Empirical LR kernel

KW - I-vectors

KW - NIST SRE

KW - Probabilistic linear discriminant analysis

KW - Relevance vector machines

UR - http://www.scopus.com/inward/record.url?scp=84912119989&partnerID=8YFLogxK

U2 - 10.1109/ISCSLP.2014.6936591

DO - 10.1109/ISCSLP.2014.6936591

M3 - Conference article published in proceeding or book

SP - 64

EP - 68

BT - Proceedings of the 9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014

PB - IEEE

T2 - 9th International Symposium on Chinese Spoken Language Processing, ISCSLP 2014

Y2 - 12 September 2014 through 14 September 2014

ER -

Relevance vector machines with empirical likelihood-ratio kernels for PLDA speaker verification

Abstract

Conference

Keywords

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this