GaussianMapping

GaussianMapping(
    self,
    mean: Tensor,
    cov: Tensor,
    reference: GaussianReference | None = None,
    diag: bool = False,
)

A mapping between two Gaussian densities.

This preconditioner provides a mapping between the standard Gaussian density and an Gaussian density with an arbitrary mean and covariance.

Parameters

mean : Tensor

The mean of the target Gaussian density.

cov : Tensor

The covariance matrix of the target Gaussian density.

reference : GaussianReference | None = None

The reference density. If this is not specified, it will be set to the unit Gaussian density with support on \([-4, 4]^{d}\).

diag : bool = False

Whether cov is a diagonal matrix.