Sobolev Norm Learning Rates for Conditional Mean Embeddings

Abstract

We develop novel learning rates for conditional mean embeddings by applying the theory of interpolation for reproducing kernel Hilbert spaces (RKHS). We derive explicit, adaptive convergence rates for the sample estimator under the misspecifed setting, where the target operator is not Hilbert-Schmidt or bounded with respect to the input/output RKHSs. We demonstrate that in certain parameter regimes, we can achieve uniform convergence rates in the output RKHS. We hope our analyses will allow the much broader application of conditional mean embeddings to more complex ML/RL settings involving infinite dimensional RKHSs and continuous state spaces.

Appears in AISTATS 2022

References in corpus (12)

• Kernel dimension reduction in regression
• Gaussian Processes and Kernel Methods: A Review on Connections and Equivalences
• Eigendecompositions of Transfer Operators in Reproducing Kernel Hilbert Spaces
• Optimal Rates for Spectral Algorithms with Least-Squares Regression over Hilbert Spaces
• Modelling transition dynamics in MDPs with RKHS embeddings
• A Rigorous Theory of Conditional Mean Embeddings
• A Measure-Theoretic Approach to Kernel Conditional Mean Embeddings
• Sobolev Norm Learning Rates for Regularized Least-Squares Algorithm
• Optimal Convergence for Distributed Learning with Stochastic Gradient Methods and Spectral Algorithms
• Conditional mean embeddings as regressors - supplementary
• Kernel Autocovariance Operators of Stationary Processes: Estimation and Convergence
• Nonparametric approximation of conditional expectation operators