Sparse Spectrum Gaussian Process Regression Articles uri icon

authors

  • LAZARO GREDILLA, MIGUEL
  • QUIÑONERO CANDELA, JOAQUIN
  • RASMUSSEN, C. E.
  • FIGUEIRAS VIDAL, ANIBAL RAMON

publication date

  • June 2010

start page

  • 1865

end page

  • 1881

volume

  • 11

International Standard Serial Number (ISSN)

  • 1532-4435

Electronic International Standard Serial Number (EISSN)

  • 1533-7928

abstract

  • We present a new sparse Gaussian Process (GP) model for regression. The key novel idea is to sparsify the spectral representation of the GP. This leads to a simple, practical algorithm for regression tasks. We compare the achievable trade-offs between predictive accuracy
    and computational requirements, and show that these are typically
    superior to existing state-of-the-art sparse approximations. We discuss
    both the weight space and function space representations, and note that
    the new construction implies priors over functions which are always
    stationary, and can approximate any covariance function in this class.