Sparse representation for Gaussian process models

Lehel Csató, Manfred Opper

    Research output: Contribution to journalArticle

    Abstract

    We develop an approach for a sparse representation for Gaussian Process (GP) models in order to overcome the limitations of GPs caused by large data sets. The method is based on a combination of a Bayesian online algorithm together with a sequential construction of a relevant subsample of the data which fully specifies the prediction of the model. Experimental results on toy examples and large real-world datasets indicate the efficiency of the approach.
    Original languageEnglish
    Pages (from-to)444-450
    Number of pages7
    JournalAdvances in Neural Information Processing Systems
    Volume13
    Publication statusPublished - 2002
    EventAdvances in Neural Information Processing Systems 1994 - Singapore, Singapore
    Duration: 16 Nov 199418 Nov 1994

    Bibliographical note

    Availble on Google books

    Keywords

    • sparse representation
    • Gaussian Process
    • limitations
    • large data sets
    • Bayesian online algorithm
    • subsample

    Cite this

    Csató, Lehel ; Opper, Manfred. / Sparse representation for Gaussian process models. In: Advances in Neural Information Processing Systems. 2002 ; Vol. 13. pp. 444-450.
    @article{667d873e1fe64b3983e87495fdb62c0e,
    title = "Sparse representation for Gaussian process models",
    abstract = "We develop an approach for a sparse representation for Gaussian Process (GP) models in order to overcome the limitations of GPs caused by large data sets. The method is based on a combination of a Bayesian online algorithm together with a sequential construction of a relevant subsample of the data which fully specifies the prediction of the model. Experimental results on toy examples and large real-world datasets indicate the efficiency of the approach.",
    keywords = "sparse representation, Gaussian Process, limitations, large data sets, Bayesian online algorithm, subsample",
    author = "Lehel Csat{\'o} and Manfred Opper",
    note = "Availble on Google books",
    year = "2002",
    language = "English",
    volume = "13",
    pages = "444--450",
    journal = "Advances in Neural Information Processing Systems",
    issn = "1049-5258",

    }

    Sparse representation for Gaussian process models. / Csató, Lehel; Opper, Manfred.

    In: Advances in Neural Information Processing Systems, Vol. 13, 2002, p. 444-450.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - Sparse representation for Gaussian process models

    AU - Csató, Lehel

    AU - Opper, Manfred

    N1 - Availble on Google books

    PY - 2002

    Y1 - 2002

    N2 - We develop an approach for a sparse representation for Gaussian Process (GP) models in order to overcome the limitations of GPs caused by large data sets. The method is based on a combination of a Bayesian online algorithm together with a sequential construction of a relevant subsample of the data which fully specifies the prediction of the model. Experimental results on toy examples and large real-world datasets indicate the efficiency of the approach.

    AB - We develop an approach for a sparse representation for Gaussian Process (GP) models in order to overcome the limitations of GPs caused by large data sets. The method is based on a combination of a Bayesian online algorithm together with a sequential construction of a relevant subsample of the data which fully specifies the prediction of the model. Experimental results on toy examples and large real-world datasets indicate the efficiency of the approach.

    KW - sparse representation

    KW - Gaussian Process

    KW - limitations

    KW - large data sets

    KW - Bayesian online algorithm

    KW - subsample

    UR - http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=8662

    M3 - Article

    VL - 13

    SP - 444

    EP - 450

    JO - Advances in Neural Information Processing Systems

    JF - Advances in Neural Information Processing Systems

    SN - 1049-5258

    ER -