English

A Measure-Theoretic Approach to Kernel Conditional Mean Embeddings

Machine Learning 2021-01-11 v8 Machine Learning

Abstract

We present an operator-free, measure-theoretic approach to the conditional mean embedding (CME) as a random variable taking values in a reproducing kernel Hilbert space. While the kernel mean embedding of unconditional distributions has been defined rigorously, the existing operator-based approach of the conditional version depends on stringent assumptions that hinder its analysis. We overcome this limitation via a measure-theoretic treatment of CMEs. We derive a natural regression interpretation to obtain empirical estimates, and provide a thorough theoretical analysis thereof, including universal consistency. As natural by-products, we obtain the conditional analogues of the maximum mean discrepancy and Hilbert-Schmidt independence criterion, and demonstrate their behaviour via simulations.

Keywords

Cite

@article{arxiv.2002.03689,
  title  = {A Measure-Theoretic Approach to Kernel Conditional Mean Embeddings},
  author = {Junhyung Park and Krikamol Muandet},
  journal= {arXiv preprint arXiv:2002.03689},
  year   = {2021}
}
R2 v1 2026-06-23T13:36:32.976Z