Generalized Singular Value Thresholding

This work studies the Generalized Singular Value Thresholding (GSVT) operator ${\Prox}_{g}^{\bm{\sigma}}(\cdot)$, \begin{equation*} {\Prox}_{g}^{\bm{\sigma}}(\B)=\arg\min\limits_{\X}\sum_{i=1}^{m}g(\sigma_{i}(\X)) + \frac{1}{2}||\X-\B||_{F}^{2}, \end{equation*} associated with a nonconvex function $g$ defined on the singular values of $\X$. We prove that GSVT can be obtained by performing the proximal operator of $g$ (denoted as $\Prox_g(\cdot)$) on the singular values since $\Prox_g(\cdot)$ is monotone when $g$ is lower bounded. If the nonconvex $g$ satisfies some conditions (many popular nonconvex surrogate functions, e.g., $\ell_p$-norm, $0Singular Value Thresholding (SVT) which is a basic subroutine in many convex low rank minimization methods. We are able to solve the nonconvex low rank minimization problem by using GSVT in place of SVT.