Abstract

We consider the problem of learning convolution operators associated tocompact Abelian groups. We study a regularization-based approach and providecorresponding learning guarantees under natural regularity conditions on theconvolution kernel. More precisely, we assume the convolution kernel is afunction in a translation invariant Hilbert space and analyze a natural ridgeregression (RR) estimator. Building on existing results for RR, we characterizethe accuracy of the estimator in terms of finite sample bounds. Interestingly,regularity assumptions which are classical in the analysis of RR, have a noveland natural interpretation in terms of space/frequency localization.Theoretical results are illustrated by numerical simulations.

 

Quick Read (beta)