Exact GPs (Regression)

Regression with a Gaussian noise model is the canonical example of Gaussian processes. These examples will work for small to medium sized datasets (~2,000 data points). All examples here use exact GP inference.

• Simple GP Regression is the basic tutorial for regression in GPyTorch.

• Spectral Mixture Regression extends on the previous example with a more complex kernel.

• Fully Bayesian GP Regression demonstrates how to perform fully Bayesian inference by sampling the GP hyperparameters using NUTS. (This example requires Pyro to be installed).

• Distributional GP Regression is an example of how to take account of uncertainty in inputs.

• Dirichlet Classification is an example of how to perform regression on classification labels via an approximate likelihood.