Automatic reproducing kernel and regularization for learning convolution kernels

TL;DR

This paper introduces a data-adaptive reproducing kernel Hilbert space (RKHS) that automatically selects kernels based on data, improving the regularization of convolution kernel learning in inverse problems.

Contribution

The authors develop a novel automatic, data-driven RKHS framework that eliminates manual kernel selection and provides efficient algorithms for convolution kernel estimation.

Findings

Automatic RKHS outperforms standard ridge regression.

The method effectively handles integral, nonlocal, and aggregation operators.

Algorithms are scalable and include Tikhonov and hybrid approaches.

Abstract

Learning convolution kernels in operators from data arises in numerous applications and represents an ill-posed inverse problem of broad interest. With scant prior information, kernel methods offer a natural nonparametric approach with regularization. However, a major challenge is to select a proper reproducing kernel, especially as operators and data vary. We show that the input data and convolution operator themselves induce an automatic, data-adaptive RKHS (DA-RKHS), obviating manual kernel selection. In particular, when the observation data is discrete and finite, there is a finite set of automatic basis functions sufficient to represent the estimators in the DA-RKHS, including the minimal-norm least-squares, Tikhonov, and conjugate-gradient estimators. We develop both Tikhonov and scalable iterative and hybrid algorithms using the automatic basis functions. Numerical experiments on integral, nonlocal, and aggregation operators confirm that our automatic RKHS regularization consistently outperforms standard ridge regression and Gaussian process methods with preselected kernels.