Nonlinear process convolutions for multioutput Gaussian processes
[edit]
Proceedings of Machine Learning Research, PMLR 89:19691977, 2019.
Abstract
The paper introduces a nonlinear version of the process convolution formalism for building covariance functions for multioutput Gaussian processes. The nonlinearity is introduced via Volterra series, one series per each output. We provide closedform expressions for the mean function and the covariance function of the approximated Gaussian process at the output of the Volterra series. The mean function and covariance function for the joint Gaussian process are derived using formulae for the product moments of Gaussian variables. We compare the performance of the nonlinear model against the classical process convolution approach in one synthetic dataset and two real datasets.
Related Material


