Efficient Multioutput Gaussian Processes through Variational Inducing Kernels

Interest in multioutput kernel methods is increasing, whether under the guise of multitask learning, multisensor networks or structured output data. From the Gaussian process perspective a multioutput Mercer kernel is a covariance function over correlated output functions. One way of constructing such kernels is based on convolution processes (CP). A key problem for this approach is efficient inference.

• Type: Journal
• Year: 2010
• Where: JMLR
• Authors: Mauricio Alvarez, David Luengo, Michalis Titsias, Neil D. Lawrence
• Keywords: Correlated Output Functions, Efficient Inference, JMLR 2010, Multioutput Kernel Methods