Non-smooth Non-convex Bregman Minimization: Unification and new Algorithms
This work provides a theoretical unification and new algorithmic framework for non-smooth non-convex optimization, benefiting researchers in optimization and applied fields by relaxing common assumptions like Lipschitz gradients.
The paper proposes a unifying algorithm for non-smooth non-convex optimization that uses convex model functions and Bregman proximal points, proving subsequential convergence to a stationary point without requiring Lipschitz continuous gradients. The algorithm generalizes several existing methods and applies to inverse problems in signal processing and machine learning.
We propose a unifying algorithm for non-smooth non-convex optimization. The algorithm approximates the objective function by a convex model function and finds an approximate (Bregman) proximal point of the convex model. This approximate minimizer of the model function yields a descent direction, along which the next iterate is found. Complemented with an Armijo-like line search strategy, we obtain a flexible algorithm for which we prove (subsequential) convergence to a stationary point under weak assumptions on the growth of the model function error. Special instances of the algorithm with a Euclidean distance function are, for example, Gradient Descent, Forward--Backward Splitting, ProxDescent, without the common requirement of a "Lipschitz continuous gradient". In addition, we consider a broad class of Bregman distance functions (generated by Legendre functions) replacing the Euclidean distance. The algorithm has a wide range of applications including many linear and non-linear inverse problems in signal/image processing and machine learning.