A0810
Title: Regularized Kaczmarz algorithm
Authors: Xuemei Chen - University of North Carolina Wilmington (United States) [presenting]
Abstract: A novel algorithmic framework based on the Kaczmarz algorithm is proposed for tensor recovery. This is an iterative algorithm for solving a linear constrained optimization problem (with a regularizer). We provide a thorough convergence analysis and its applications from the vector case to the tensor one. We also show numerical results on a variety of tensor recovery applications to illustrate the enormous potential of the proposed methods.