Recovering Distributions from Gaussian RKHS Embeddings

Motonobu Kanagawa, Kenji Fukumizu

Abstract

Recent advances of kernel methods have yielded a framework for nonparametric statistical inference called RKHS embeddings, in which all probability distributions are represented as elements in a re...
BibTeX
@inproceedings{kanagawa2014recovering,
  author = {Motonobu Kanagawa and Kenji Fukumizu},
  title = {{Recovering Distributions from Gaussian RKHS Embeddings}},
  booktitle = {International Conference on Artificial Intelligence and Statistics},
  year = {2014},
}
Figure