Inter-domain Deep Gaussian Processes

Tim G. J. Rudner, Dino Sejdinovic, Yarin Gal
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:8286-8294, 2020.

Abstract

Inter-domain Gaussian processes (GPs) allow for high flexibility and low computational cost when performing approximate inference in GP models. They are particularly suitable for modeling data exhibiting global structure but are limited to stationary covariance functions and thus fail to model non-stationary data effectively. We propose Inter-domain Deep Gaussian Processes, an extension of inter-domain shallow GPs that combines the advantages of inter-domain and deep Gaussian processes (DGPs), and demonstrate how to leverage existing approximate inference methods to perform simple and scalable approximate inference using inter-domain features in DGPs. We assess the performance of our method on a range of regression tasks and demonstrate that it outperforms inter-domain shallow GPs and conventional DGPs on challenging large-scale real-world datasets exhibiting both global structure as well as a high-degree of non-stationarity.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-rudner20a, title = {Inter-domain Deep {G}aussian Processes}, author = {Rudner, Tim G. J. and Sejdinovic, Dino and Gal, Yarin}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {8286--8294}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/rudner20a/rudner20a.pdf}, url = {https://proceedings.mlr.press/v119/rudner20a.html}, abstract = {Inter-domain Gaussian processes (GPs) allow for high flexibility and low computational cost when performing approximate inference in GP models. They are particularly suitable for modeling data exhibiting global structure but are limited to stationary covariance functions and thus fail to model non-stationary data effectively. We propose Inter-domain Deep Gaussian Processes, an extension of inter-domain shallow GPs that combines the advantages of inter-domain and deep Gaussian processes (DGPs), and demonstrate how to leverage existing approximate inference methods to perform simple and scalable approximate inference using inter-domain features in DGPs. We assess the performance of our method on a range of regression tasks and demonstrate that it outperforms inter-domain shallow GPs and conventional DGPs on challenging large-scale real-world datasets exhibiting both global structure as well as a high-degree of non-stationarity.} }
Endnote
%0 Conference Paper %T Inter-domain Deep Gaussian Processes %A Tim G. J. Rudner %A Dino Sejdinovic %A Yarin Gal %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-rudner20a %I PMLR %P 8286--8294 %U https://proceedings.mlr.press/v119/rudner20a.html %V 119 %X Inter-domain Gaussian processes (GPs) allow for high flexibility and low computational cost when performing approximate inference in GP models. They are particularly suitable for modeling data exhibiting global structure but are limited to stationary covariance functions and thus fail to model non-stationary data effectively. We propose Inter-domain Deep Gaussian Processes, an extension of inter-domain shallow GPs that combines the advantages of inter-domain and deep Gaussian processes (DGPs), and demonstrate how to leverage existing approximate inference methods to perform simple and scalable approximate inference using inter-domain features in DGPs. We assess the performance of our method on a range of regression tasks and demonstrate that it outperforms inter-domain shallow GPs and conventional DGPs on challenging large-scale real-world datasets exhibiting both global structure as well as a high-degree of non-stationarity.
APA
Rudner, T.G.J., Sejdinovic, D. & Gal, Y.. (2020). Inter-domain Deep Gaussian Processes. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:8286-8294 Available from https://proceedings.mlr.press/v119/rudner20a.html.

Related Material