Timezone: »

Spatio-Temporal Variational Gaussian Processes
Oliver Hamelijnck · William Wilkinson · Niki Loppi · Arno Solin · Theodoros Damoulas

Tue Dec 07 08:30 AM -- 10:00 AM (PST) @ None #None

We introduce a scalable approach to Gaussian process inference that combines spatio-temporal filtering with natural gradient variational inference, resulting in a non-conjugate GP method for multivariate data that scales linearly with respect to time. Our natural gradient approach enables application of parallel filtering and smoothing, further reducing the temporal span complexity to be logarithmic in the number of time steps. We derive a sparse approximation that constructs a state-space model over a reduced set of spatial inducing points, and show that for separable Markov kernels the full and sparse cases exactly recover the standard variational GP, whilst exhibiting favourable computational properties. To further improve the spatial scaling we propose a mean-field assumption of independence between spatial locations which, when coupled with sparsity and parallelisation, leads to an efficient and accurate method for large spatio-temporal problems.

Author Information

Oliver Hamelijnck (The Alan Turing Institute)
William Wilkinson (Aalto University)
Niki Loppi (NVIDIA)
Arno Solin (Aalto University)
Theodoros Damoulas (University of Warwick)

More from the Same Authors