Sobolev Norm Learning Rates for Conditional Mean Embeddings
share
We develop novel learning rates for conditional mean embeddings by applying the theory of interpolation for reproducing kernel Hilbert spaces (RKHS). Our learning rates demonstrate consistency of the sample estimator under drastically weaker assumptions than the state-of-the art, allowing the much broader application of conditional mean embeddings to more complex ML/RL settings involving infinite dimensional RKHS and continuous state spaces.
READ FULL TEXT