Speaker recognition using multiple kernel learning based on conditional entropy minimization

Tetsuji Ogawa, Hideitsu Hino, Nima Reyhani, Noboru Murata, Tetsunori Kobayashi

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    4 Citations (Scopus)

    Abstract

    We applied a multiple kernel learning (MKL) method based on information-theoretic optimization to speaker recognition. Most of the kernel methods applied to speaker recognition systems require a suitable kernel function and its parameters to be determined for a given data set. In contrast, MKL eliminates the need for strict determination of the kernel function and parameters by using a convex combination of element kernels. In the present paper, we describe an MKL algorithm based on conditional entropy minimization (MCEM). We experimentally verified the effectiveness of MCEM for speaker classification; this method reduced the speaker error rate as compared to conventional methods.

    Original languageEnglish
    Title of host publicationICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
    Pages2204-2207
    Number of pages4
    DOIs
    Publication statusPublished - 2011
    Event36th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2011 - Prague
    Duration: 2011 May 222011 May 27

    Other

    Other36th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2011
    CityPrague
    Period11/5/2211/5/27

    Fingerprint

    Entropy
    Learning algorithms

    Keywords

    • MCEM
    • Multiple kernel learning
    • speaker recognition

    ASJC Scopus subject areas

    • Signal Processing
    • Software
    • Electrical and Electronic Engineering

    Cite this

    Ogawa, T., Hino, H., Reyhani, N., Murata, N., & Kobayashi, T. (2011). Speaker recognition using multiple kernel learning based on conditional entropy minimization. In ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings (pp. 2204-2207). [5946918] https://doi.org/10.1109/ICASSP.2011.5946918

    Speaker recognition using multiple kernel learning based on conditional entropy minimization. / Ogawa, Tetsuji; Hino, Hideitsu; Reyhani, Nima; Murata, Noboru; Kobayashi, Tetsunori.

    ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. 2011. p. 2204-2207 5946918.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Ogawa, T, Hino, H, Reyhani, N, Murata, N & Kobayashi, T 2011, Speaker recognition using multiple kernel learning based on conditional entropy minimization. in ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings., 5946918, pp. 2204-2207, 36th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2011, Prague, 11/5/22. https://doi.org/10.1109/ICASSP.2011.5946918
    Ogawa T, Hino H, Reyhani N, Murata N, Kobayashi T. Speaker recognition using multiple kernel learning based on conditional entropy minimization. In ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. 2011. p. 2204-2207. 5946918 https://doi.org/10.1109/ICASSP.2011.5946918
    Ogawa, Tetsuji ; Hino, Hideitsu ; Reyhani, Nima ; Murata, Noboru ; Kobayashi, Tetsunori. / Speaker recognition using multiple kernel learning based on conditional entropy minimization. ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. 2011. pp. 2204-2207
    @inproceedings{abf9777160824307900cda7f6e712e60,
    title = "Speaker recognition using multiple kernel learning based on conditional entropy minimization",
    abstract = "We applied a multiple kernel learning (MKL) method based on information-theoretic optimization to speaker recognition. Most of the kernel methods applied to speaker recognition systems require a suitable kernel function and its parameters to be determined for a given data set. In contrast, MKL eliminates the need for strict determination of the kernel function and parameters by using a convex combination of element kernels. In the present paper, we describe an MKL algorithm based on conditional entropy minimization (MCEM). We experimentally verified the effectiveness of MCEM for speaker classification; this method reduced the speaker error rate as compared to conventional methods.",
    keywords = "MCEM, Multiple kernel learning, speaker recognition",
    author = "Tetsuji Ogawa and Hideitsu Hino and Nima Reyhani and Noboru Murata and Tetsunori Kobayashi",
    year = "2011",
    doi = "10.1109/ICASSP.2011.5946918",
    language = "English",
    isbn = "9781457705397",
    pages = "2204--2207",
    booktitle = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

    }

    TY - GEN

    T1 - Speaker recognition using multiple kernel learning based on conditional entropy minimization

    AU - Ogawa, Tetsuji

    AU - Hino, Hideitsu

    AU - Reyhani, Nima

    AU - Murata, Noboru

    AU - Kobayashi, Tetsunori

    PY - 2011

    Y1 - 2011

    N2 - We applied a multiple kernel learning (MKL) method based on information-theoretic optimization to speaker recognition. Most of the kernel methods applied to speaker recognition systems require a suitable kernel function and its parameters to be determined for a given data set. In contrast, MKL eliminates the need for strict determination of the kernel function and parameters by using a convex combination of element kernels. In the present paper, we describe an MKL algorithm based on conditional entropy minimization (MCEM). We experimentally verified the effectiveness of MCEM for speaker classification; this method reduced the speaker error rate as compared to conventional methods.

    AB - We applied a multiple kernel learning (MKL) method based on information-theoretic optimization to speaker recognition. Most of the kernel methods applied to speaker recognition systems require a suitable kernel function and its parameters to be determined for a given data set. In contrast, MKL eliminates the need for strict determination of the kernel function and parameters by using a convex combination of element kernels. In the present paper, we describe an MKL algorithm based on conditional entropy minimization (MCEM). We experimentally verified the effectiveness of MCEM for speaker classification; this method reduced the speaker error rate as compared to conventional methods.

    KW - MCEM

    KW - Multiple kernel learning

    KW - speaker recognition

    UR - http://www.scopus.com/inward/record.url?scp=80051657839&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=80051657839&partnerID=8YFLogxK

    U2 - 10.1109/ICASSP.2011.5946918

    DO - 10.1109/ICASSP.2011.5946918

    M3 - Conference contribution

    AN - SCOPUS:80051657839

    SN - 9781457705397

    SP - 2204

    EP - 2207

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

    ER -