Efficient training of Gaussian processes with tensor product structure

Josie König, Max Pfeffer, Martin Stoll

To determine the optimal set of hyperparameters of a Gaussian process based on a large number of training data, both a linear system and a trace estimation problem must be solved.…