WorksMathematics
Inversion of Set-Valued Wiener Models
Abstract
A Wiener system consists of a linear system followed by a nonlinear system. This paper addresses the problem of recovering a vector in a Hilbert space from the outputs of multiple Wiener systems. While prior literature on provably convergent methods has studied this problem under the assumption that the nonlinear operators are single-valued and firmly nonexpansive, we investigate the broader setting in which the nonlinear operators are set-valued and maximally cohypomonotone. We design a splitting algorithm for this extended framework and establish its convergence. Our framework remains novel even in the special case of maximally monotone operators. Numerical applications to signal and image recovery are presented.