Sparse Spectrum Gaussian Process Regression

被引：0

作者：

Lazaro-Gredilla, Miguel ^{[1
]}

Quinonero-Candela, Joaquin ^{[3
]}

Rasmussen, Carl Edward ^{[2
,4
]}

Figueiras-Vidal, Anibal R. ^{[1
]}

机构：

[1] Univ Carlos III Madrid, Dept Teoria Senal & Comunicac, Madrid 28911, Spain

[2] Max Planck Inst Biol Cybernet, D-72076 Tubingen, Germany

[3] Microsoft Res Ltd, Cambridge CB3 0FB, England

[4] Univ Cambridge, Dept Engn, Cambridge CB2 1PZ, England

来源：

JOURNAL OF MACHINE LEARNING RESEARCH | 2010年 / 11卷

关键词：

Gaussian process; probabilistic regression; sparse approximation; power spectrum; computational efficiency;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

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.

引用

页码：1865 / 1881

页数：17