Sparse Convolved Gaussian Processes for Multi-output Regression

Mauricio A. ÁlvarezNeil D. Lawrence
Advances in Neural Information Processing Systems, MIT Press 21:57-64, 2008.

Abstract

We present a sparse approximation approach for dependent output Gaussian processes (GP). Employing a latent function framework, we apply the convolution process formalism to establish dependencies between output variables, where each latent function is represented as a GP. Based on these latent functions, we establish an approximation scheme using a conditional independence assumption between the output processes, leading to an approximation of the full covariance which is determined by the locations at which the latent functions are evaluated. We show results of the proposed methodology for synthetic data and real world applications on pollution prediction and a sensor network.

Cite this Paper


BibTeX
@InProceedings{Alvarez:convolved08, title = {Sparse Convolved {G}aussian Processes for Multi-output Regression}, author = {Álvarez, Mauricio A. and Lawrence, Neil D.}, booktitle = {Advances in Neural Information Processing Systems}, pages = {57--64}, year = {2008}, editor = {Koller, Daphne and Schuurmans, Dale and Bengio, Yoshua and Bottou, Leon}, volume = {21}, address = {Cambridge, MA}, publisher = {MIT Press}, pdf = {https://papers.nips.cc/paper/2008/file/149e9677a5989fd342ae44213df68868-Paper.pdf}, url = {http://inverseprobability.com/publications/alvarez-convolved08.html}, abstract = {We present a sparse approximation approach for dependent output Gaussian processes (GP). Employing a latent function framework, we apply the convolution process formalism to establish dependencies between output variables, where each latent function is represented as a GP. Based on these latent functions, we establish an approximation scheme using a conditional independence assumption between the output processes, leading to an approximation of the full covariance which is determined by the locations at which the latent functions are evaluated. We show results of the proposed methodology for synthetic data and real world applications on pollution prediction and a sensor network.} }
Endnote
%0 Conference Paper %T Sparse Convolved Gaussian Processes for Multi-output Regression %A Mauricio A. Álvarez %A Neil D. Lawrence %B Advances in Neural Information Processing Systems %D 2008 %E Daphne Koller %E Dale Schuurmans %E Yoshua Bengio %E Leon Bottou %F Alvarez:convolved08 %I MIT Press %P 57--64 %U http://inverseprobability.com/publications/alvarez-convolved08.html %V 21 %X We present a sparse approximation approach for dependent output Gaussian processes (GP). Employing a latent function framework, we apply the convolution process formalism to establish dependencies between output variables, where each latent function is represented as a GP. Based on these latent functions, we establish an approximation scheme using a conditional independence assumption between the output processes, leading to an approximation of the full covariance which is determined by the locations at which the latent functions are evaluated. We show results of the proposed methodology for synthetic data and real world applications on pollution prediction and a sensor network.
RIS
TY - CPAPER TI - Sparse Convolved Gaussian Processes for Multi-output Regression AU - Mauricio A. Álvarez AU - Neil D. Lawrence BT - Advances in Neural Information Processing Systems DA - 2008/12/08 ED - Daphne Koller ED - Dale Schuurmans ED - Yoshua Bengio ED - Leon Bottou ID - Alvarez:convolved08 PB - MIT Press VL - 21 SP - 57 EP - 64 L1 - https://papers.nips.cc/paper/2008/file/149e9677a5989fd342ae44213df68868-Paper.pdf UR - http://inverseprobability.com/publications/alvarez-convolved08.html AB - We present a sparse approximation approach for dependent output Gaussian processes (GP). Employing a latent function framework, we apply the convolution process formalism to establish dependencies between output variables, where each latent function is represented as a GP. Based on these latent functions, we establish an approximation scheme using a conditional independence assumption between the output processes, leading to an approximation of the full covariance which is determined by the locations at which the latent functions are evaluated. We show results of the proposed methodology for synthetic data and real world applications on pollution prediction and a sensor network. ER -
APA
Álvarez, M.A. & Lawrence, N.D.. (2008). Sparse Convolved Gaussian Processes for Multi-output Regression. Advances in Neural Information Processing Systems 21:57-64 Available from http://inverseprobability.com/publications/alvarez-convolved08.html.

Related Material