Iterative regularization via dual diagonal descent

Abstract

In the context of linear inverse problems, we propose and study a general iterative regularization method allowing to consider large classes of regularizers and data-fit terms. The algorithm we propose is based on a primal-dual diagonal {descent} method. Our analysis establishes convergence as well as stability results. Theoretical findings are complemented with numerical experiments showing state of the art performances.

41 pages, 13 figures. 4-pages version of the paper available at http://opt-ml.org/papers/OPT2016_paper_19.pdf

References in corpus (4)

• Don't relax: early stopping for convex regularization
• Splitting forward-backward penalty scheme for constrained variational problems
• Infimal convolution of data discrepancies for mixed noise removal
• Asymptotic behavior of nonautonomous monotone and subgradient evolution equations

Cited by in corpus (6)

• Accelerated iterative regularization via dual diagonal descent
• An iterative regularized mirror descent method for ill-posed nondifferentiable stochastic optimization
• The primal-dual hybrid gradient method reduces to a primal method for linearly constrained optimization problems
• Iterative regularization for convex regularizers
• A Randomized Block Coordinate Iterative Regularized Gradient Method for High-dimensional Ill-posed Convex Optimization
• New Pair of Primal Dual Algorithms for Bregman Iterated Variational Regularization