Gaussian Process Random Fields
This work addresses scalability issues in Gaussian processes for practitioners dealing with large datasets, though it is incremental as it builds on existing GP approximations.
The authors tackled the computational complexity of Gaussian processes by introducing Gaussian Process Random Fields (GPRF), which approximate large-scale GPs using local GPs coupled via pairwise potentials, enabling latent variable modeling and hyperparameter selection on large datasets, as demonstrated on synthetic spatial data and a real-world seismic event location application.
Gaussian processes have been successful in both supervised and unsupervised machine learning tasks, but their computational complexity has constrained practical applications. We introduce a new approximation for large-scale Gaussian processes, the Gaussian Process Random Field (GPRF), in which local GPs are coupled via pairwise potentials. The GPRF likelihood is a simple, tractable, and parallelizeable approximation to the full GP marginal likelihood, enabling latent variable modeling and hyperparameter selection on large datasets. We demonstrate its effectiveness on synthetic spatial data as well as a real-world application to seismic event location.