KernelHerding.KernelHerdingGradient — TypeThe gradient < x - μ, . > is represented by x and μ.
KernelHerding.KernelHerdingIterate — TypeKernel herding iterate.
KernelHerding.MarginalPolytopeWahba — TypeThe marginal polytope of the Wahba kernel.
KernelHerding.MeanElement — TypeMeanElement μ must implement dot with a functional.
KernelHerding.NonZeroMeanElement — Typeμ =/= 0.
KernelHerding.ZeroMeanElement — Typeμ = 0.
Base.:* — MethodMultiplication for KernelHerdingIterate.
Base.:* — MethodMultiplication for KernelHerdingIterate, different order.
Base.:+ — MethodAddition for KernelHerdingIterate.
Base.:- — MethodSubtraction for KernelHerdingIterate.
FrankWolfe.compute_extreme_point — MethodComputes the extreme point in the Frank-Wolfe algorithm for kernel herding in Wahba's kernel.
KernelHerding.bernoulli_polynomial — MethodThe degree-2 Bernoulli polynomial.
KernelHerding.construct_rho — MethodGiven rho =(rhoa, rhob), returns a normalizedrho such that normalizedrho is a valid distribution.
KernelHerding.create_loss_function_gradient — MethodCreates the loss function and the gradient function, given a MeanElement μ, that is, 1/2 || x - μ ||_H² and < x - μ, . >, respectively.
KernelHerding.kernel_evaluation_wahba — MethodEvaluates the Wahba kernel over two real numbers.
KernelHerding.mu_from_rho — MethodGiven a distribution ρ, computes μ.
KernelHerding.pad_non_zero_mean_element! — MethodThe NonZeroMeanElement is represented by two vectors, cosineweights and sineweights. This function pads the shorter vector with zeros.
LinearAlgebra.dot — MethodScalar product for KernelHerdingIterate with KernelHerdingGradient.
LinearAlgebra.dot — MethodScalar product for two KernelHerdingIterates.
LinearAlgebra.dot — MethodScalar product for KernelHerdingIterate with NonZeroMeanElement.
LinearAlgebra.dot — MethodScalar product for KernelHerdingIterate with ZeroMeanElement.
LinearAlgebra.norm — MethodNorm of NonZeroMeanElement, corresponds to ||µ||.
LinearAlgebra.norm — MethodNorm of ZeroMeanElement, corresponds to ||µ||.