Parallelizable sparse inverse formulation Gaussian processes (SpInGP)

We propose a parallelizable sparse inverse formulation Gaussian process (SpInGP) al- gorithm for temporal Gaussian process mod- els. It uses a sparse precision GP formulation and sparse matrix routines to speed up the computations. Due to the state-space formu- lation used in the algorithm, the time com- plexity of the basic SpInGP is linear, and be- cause all the computations are parallelizable, the parallel form of the algorithm is sublin- ear in the number of data points. We provide example algorithms to implement the sparse matrix routines and experimentally test the method using both simulated and real data.