Lloyd's algorithm uses iterative improvements to find a locally optimalk-means clustering. The performance of Lloyd'salgorithm crucially depends on the quality of the initial clustering, which isdefined bytheinitial setofcenters, called aseed.

In this work, we provide evidence to the contrary.

In this work, we propose a Bayesian formulation of deconditioning which naturally recovers the initial reproducing kernel Hilbert space formulation from Hsu and Ramos[1]. We extend deconditioning to a downscaling setup and devise efficient conditional mean embedding estimator for multiresolution data.