Stein Unbiased GrAdient estimator of the Risk (SUGAR) for multiple parameter selection

TL;DR

This paper introduces SUGAR, a gradient estimator for the risk function in models with non-smooth regularizers, enabling efficient parameter optimization in inverse problems.

Contribution

It develops a weakly differentiable approximation of SURE called SUGAR, allowing gradient-based optimization for models with non-smooth regularizers like l1-norm.

Findings

SUGAR provides an asymptotically unbiased gradient estimate.

It is consistent for soft-thresholding regularizers.

Applications include image restoration and matrix completion.

Abstract

Algorithms to solve variational regularization of ill-posed inverse problems usually involve operators that depend on a collection of continuous parameters. When these operators enjoy some (local) regularity, these parameters can be selected using the so-called Stein Unbiased Risk Estimate (SURE). While this selection is usually performed by exhaustive search, we address in this work the problem of using the SURE to efficiently optimize for a collection of continuous parameters of the model. When considering non-smooth regularizers, such as the popular l1-norm corresponding to soft-thresholding mapping, the SURE is a discontinuous function of the parameters preventing the use of gradient descent optimization techniques. Instead, we focus on an approximation of the SURE based on finite differences as proposed in (Ramani et al., 2008). Under mild assumptions on the estimation mapping, we show that this approximation is a weakly differentiable function of the parameters and its weak gradient, coined the Stein Unbiased GrAdient estimator of the Risk (SUGAR), provides an asymptotically (with respect to the data dimension) unbiased estimate of the gradient of the risk. Moreover, in the particular case of soft-thresholding, it is proved to be also a consistent estimator. This gradient estimate can then be used as a basis to perform a quasi-Newton optimization. The computation of the SUGAR relies on the closed-form (weak) differentiation of the non-smooth function. We provide its expression for a large class of iterative methods including proximal splitting ones and apply our strategy to regularizations involving non-smooth convex structured penalties. Illustrations on various image restoration and matrix completion problems are given.