Scalable Gaussian Process Inference with Finite-data Mean and Variance Guarantees

Jonathan H. Huggins, Trevor Campbell, Mikolaj Kasprzak, Tamara Broderick
Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics, PMLR 89:796-805, 2019.

Abstract

Gaussian processes (GPs) offer a flexible class of priors for nonparametric Bayesian regression, but popular GP posterior inference methods are typically prohibitively slow or lack desirable finite-data guarantees on quality. We develop a scalable approach to approximate GP regression, with finite-data guarantees on the accuracy of our pointwise posterior mean and variance estimates. Our main contribution is a novel objective for approximate inference in the nonparametric setting: the preconditioned Fisher (pF) divergence. We show that unlike the Kullback–Leibler divergence (used in variational inference), the pF divergence bounds bounds the 2-Wasserstein distance, which in turn provides tight bounds on the pointwise error of mean and variance estimates. We demonstrate that, for sparse GP likelihood approximations, we can minimize the pF divergence bounds efficiently. Our experiments show that optimizing the pF divergence bounds has the same computational requirements as variational sparse GPs while providing comparable empirical performance—in addition to our novel finite-data quality guarantees.

Cite this Paper


BibTeX
@InProceedings{pmlr-v89-huggins19a, title = {Scalable Gaussian Process Inference with Finite-data Mean and Variance Guarantees}, author = {Huggins, Jonathan H. and Campbell, Trevor and Kasprzak, Mikolaj and Broderick, Tamara}, booktitle = {Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics}, pages = {796--805}, year = {2019}, editor = {Chaudhuri, Kamalika and Sugiyama, Masashi}, volume = {89}, series = {Proceedings of Machine Learning Research}, month = {16--18 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v89/huggins19a/huggins19a.pdf}, url = {http://proceedings.mlr.press/v89/huggins19a.html}, abstract = {Gaussian processes (GPs) offer a flexible class of priors for nonparametric Bayesian regression, but popular GP posterior inference methods are typically prohibitively slow or lack desirable finite-data guarantees on quality. We develop a scalable approach to approximate GP regression, with finite-data guarantees on the accuracy of our pointwise posterior mean and variance estimates. Our main contribution is a novel objective for approximate inference in the nonparametric setting: the preconditioned Fisher (pF) divergence. We show that unlike the Kullback–Leibler divergence (used in variational inference), the pF divergence bounds bounds the 2-Wasserstein distance, which in turn provides tight bounds on the pointwise error of mean and variance estimates. We demonstrate that, for sparse GP likelihood approximations, we can minimize the pF divergence bounds efficiently. Our experiments show that optimizing the pF divergence bounds has the same computational requirements as variational sparse GPs while providing comparable empirical performance—in addition to our novel finite-data quality guarantees.} }
Endnote
%0 Conference Paper %T Scalable Gaussian Process Inference with Finite-data Mean and Variance Guarantees %A Jonathan H. Huggins %A Trevor Campbell %A Mikolaj Kasprzak %A Tamara Broderick %B Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Masashi Sugiyama %F pmlr-v89-huggins19a %I PMLR %P 796--805 %U http://proceedings.mlr.press/v89/huggins19a.html %V 89 %X Gaussian processes (GPs) offer a flexible class of priors for nonparametric Bayesian regression, but popular GP posterior inference methods are typically prohibitively slow or lack desirable finite-data guarantees on quality. We develop a scalable approach to approximate GP regression, with finite-data guarantees on the accuracy of our pointwise posterior mean and variance estimates. Our main contribution is a novel objective for approximate inference in the nonparametric setting: the preconditioned Fisher (pF) divergence. We show that unlike the Kullback–Leibler divergence (used in variational inference), the pF divergence bounds bounds the 2-Wasserstein distance, which in turn provides tight bounds on the pointwise error of mean and variance estimates. We demonstrate that, for sparse GP likelihood approximations, we can minimize the pF divergence bounds efficiently. Our experiments show that optimizing the pF divergence bounds has the same computational requirements as variational sparse GPs while providing comparable empirical performance—in addition to our novel finite-data quality guarantees.
APA
Huggins, J.H., Campbell, T., Kasprzak, M. & Broderick, T.. (2019). Scalable Gaussian Process Inference with Finite-data Mean and Variance Guarantees. Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 89:796-805 Available from http://proceedings.mlr.press/v89/huggins19a.html.

Related Material