TY - CONF
T1 - Non-linear process convolutions for multi-output Gaussian processes
AU - Álvarez, Mauricio A.
AU - Ward, Wil O.C.
AU - Guarnizo, Cristian
N1 - Funding Information:
MAA and WW have been financed by the Engineering and Physical Research Council (EPSRC) Research Project EP/N014162/1. MAA has also been financed by the EPSRC Research Project EP/R034303/1. CG would like to thank Convocatoria 567 from Administrative Department of Science, Technology and Innovation of Colombia (COLCIENCIAS) for the financial support.
Funding Information:
MAA and WW have been financed by the Engineering and Physical Research Council (EPSRC) Research Project EP/N014162/1. MAA has also been financed by the EP-SRC Research Project EP/R034303/1. CG would like to thank Convocatoria 567 from Administrative Department of Science, Technology and Innovation of Colombia (COL-CIENCIAS) for the financial support.
Publisher Copyright:
© 2019 by the author(s).
PY - 2020
Y1 - 2020
N2 - The paper introduces a non-linear version of the process convolution formalism for building covariance functions for multi-output Gaussian processes. The non-linearity is introduced via Volterra series, one series per each output. We provide closed-form 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 non-linear model against the classical process convolution approach in one synthetic dataset and two real datasets.
AB - The paper introduces a non-linear version of the process convolution formalism for building covariance functions for multi-output Gaussian processes. The non-linearity is introduced via Volterra series, one series per each output. We provide closed-form 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 non-linear model against the classical process convolution approach in one synthetic dataset and two real datasets.
UR - http://www.scopus.com/inward/record.url?scp=85085024522&partnerID=8YFLogxK
M3 - Paper
AN - SCOPUS:85085024522
T2 - 22nd International Conference on Artificial Intelligence and Statistics, AISTATS 2019
Y2 - 16 April 2019 through 18 April 2019
ER -