Convergence and finite sample approximations of entropic regularized Wasserstein distances in Gaussian and RKHS settings.
Minh Ha QuangPublished in: CoRR (2021)
Keyphrases
- finite sample
- uniform convergence
- reproducing kernel hilbert space
- regularized least squares
- gaussian kernel
- distance measure
- sample size
- gaussian kernels
- risk minimization
- nearest neighbor
- loss function
- statistical learning theory
- error bounds
- kernel methods
- parzen window
- learning theory
- generalization bounds
- distance function
- euclidean space
- real valued
- density estimation
- special case
- generalization error
- kernel function
- least squares
- domain adaptation
- euclidean distance
- linear model
- kernel machines
- support vector
- active learning
- maximum likelihood
- gaussian process
- learning problems
- input space
- model selection
- supervised learning
- information theory
- feature vectors
- pairwise
- feature selection