Sparse Kernel Gaussian Processes through Iterative Charted Refinement (ICR). (arXiv:2206.10634v1 [cs.LG])

Web: http://arxiv.org/abs/2206.10634

Gaussian Processes (GPs) are highly expressive, probabilistic models. A major
limitation is their computational complexity. Naively, exact GP inference
requires $\mathcal{O}(N^3)$ computations with $N$ denoting the number of
modeled points. Current approaches to overcome this limitation either rely on
sparse, structured or stochastic representations of data or kernel respectively
and usually involve nested optimizations to evaluate a GP. We present a new,
generative method named Iterative Charted Refinement (ICR) to model GPs on
nearly arbitrarily spaced points in $\mathcal{O}(N)$ time …

arxiv gaussian processes iterative kernel lg processes