Vanilla Gaussian GP Controller

The GaussianGPController provides the user with a standard GP model with no extra features.

class vanguard.vanilla.GaussianGPController(train_x, train_y, kernel_class, y_std, mean_class=<class 'gpytorch.means.constant_mean.ConstantMean'>, likelihood_class=<class 'gpytorch.likelihoods.gaussian_likelihood.FixedNoiseGaussianLikelihood'>, marginal_log_likelihood_class=<class 'gpytorch.mlls.exact_marginal_log_likelihood.ExactMarginalLogLikelihood'>, optimiser_class=<class 'torch.optim.adam.Adam'>, smart_optimiser_class=<class 'vanguard.optimise.optimiser.GreedySmartOptimiser'>, rng=None, **kwargs)

Bases: GPController

Base class for implementing standard GP regression with flexible prior kernel and mean functions.

This is the best starting point for users, containing many sensible default values. The standard reference is [Rasmussen06].

Parameters:

• train_x (Union[Tensor, ndarray[tuple[Any, ...], dtype[floating]]]) – (n_samples, n_features) The inputs (or the observed values).

• train_y (Union[Tensor, ndarray[tuple[Any, ...], dtype[floating]], ndarray[tuple[Any, ...], dtype[integer]]]) – (n_samples,) or (n_samples, 1) The responsive values.

• kernel_class (type[Kernel]) – An uninstantiated subclass of gpytorch.kernels.Kernel.

• y_std (Union[Tensor, ndarray[tuple[Any, ...], dtype[floating]], float]) –

  The observation noise standard deviation, one of:

  • ndarray (n_samples,): known heteroskedastic noise.

  • float: known homoskedastic noise assumed.

• mean_class (type[Mean]) – An uninstantiated subclass of gpytorch.means.Mean to use in the prior GP. Defaults to gpytorch.means.ConstantMean.

• likelihood_class (type[Likelihood]) – An uninstantiated subclass of gpytorch.likelihoods.Likelihood. The default is gpytorch.likelihoods.FixedNoiseGaussianLikelihood.

• marginal_log_likelihood_class (type[MarginalLogLikelihood]) – An uninstantiated subclass of of an MLL from gpytorch.mlls. The default is gpytorch.mlls.ExactMarginalLogLikelihood.

• optimiser_class (type[Optimizer]) – An uninstantiated torch.optim.Optimizer class used for gradient-based learning of hyperparameters. The default is torch.optim.Adam.

• smart_optimiser_class (type[SmartOptimiser]) – An uninstantiated SmartOptimiser class used to wrap the optimiser_class and enable early stopping.

• rng (Optional[Generator]) – Generator instance used to generate random numbers.

• kwargs (Any) – For a complete list, see GPController.

__init__(train_x, train_y, kernel_class, y_std, mean_class=<class 'gpytorch.means.constant_mean.ConstantMean'>, likelihood_class=<class 'gpytorch.likelihoods.gaussian_likelihood.FixedNoiseGaussianLikelihood'>, marginal_log_likelihood_class=<class 'gpytorch.mlls.exact_marginal_log_likelihood.ExactMarginalLogLikelihood'>, optimiser_class=<class 'torch.optim.adam.Adam'>, smart_optimiser_class=<class 'vanguard.optimise.optimiser.GreedySmartOptimiser'>, rng=None, **kwargs)

Initialise self.

Parameters:

• train_x (Union[Tensor, ndarray[tuple[Any, ...], dtype[floating]]])

• train_y (Union[Tensor, ndarray[tuple[Any, ...], dtype[floating]], ndarray[tuple[Any, ...], dtype[integer]]])

• kernel_class (type[Kernel])

• y_std (Union[Tensor, ndarray[tuple[Any, ...], dtype[floating]], float])

• mean_class (type[Mean])

• likelihood_class (type[Likelihood])

• marginal_log_likelihood_class (type[MarginalLogLikelihood])

• optimiser_class (type[Optimizer])

• smart_optimiser_class (type[SmartOptimiser])

• rng (Optional[Generator])

• kwargs (Any)