Differentiating Nonsmooth Solutions to Parametric Monotone Inclusion Problems

DOI10.1137/22M1541630zbMATH Open1530.49026arXiv2212.07844OpenAlexW4311599212MaRDI QIDQ6136656FDOQ6136656

Authors: Jérôme Bolte, Edouard Pauwels, Antonio Silveti-Falls 

Publication date: 17 January 2024

Published in: SIAM Journal on Optimization (Search for Journal in Brave)

Abstract: We leverage path differentiability and a recent result on nonsmooth implicit differentiation calculus to give sufficient conditions ensuring that the solution to a monotone inclusion problem will be path differentiable, with formulas for computing its generalized gradient. A direct consequence of our result is that these solutions happen to be differentiable almost everywhere. Our approach is fully compatible with automatic differentiation and comes with assumptions which are easy to check, roughly speaking: semialgebraicity and strong monotonicity. We illustrate the scope of our results by considering three fundamental composite problem settings: strongly convex problems, dual solutions to convex minimization problems and primal-dual solutions to min-max problems.

Full work available at URL: https://arxiv.org/abs/2212.07844

Recommendations

• Sensitivity analysis of maximally monotone inclusions via the proto-differentiability of the resolvent operator
• Uniqueness and differentiability of solutions of parametric nonlinear complementarity problems
• Path differentiability of ODE flows
• scientific article; zbMATH DE number 1099074
• Conservative set valued fields, automatic differentiation, stochastic gradient methods and deep learning

zbMATH Keywords

generalized equationmaximal monotone operatormonotone inclusionClarke subdifferentialgeneralized gradientimplicit differentiationconservative fielddifferentiating solutions

Mathematics Subject Classification ID

Nonsmooth analysis (49J52) Set-valued and variational analysis (49J53) Optimality conditions for minimax problems (49K35) Sensitivity, stability, well-posedness (49K40)

Cites Work

• Splitting Algorithms for the Sum of Two Nonlinear Operators
• Variational Analysis
• Convex analysis and monotone operator theory in Hilbert spaces
• First-order methods in optimization
• Title not available (Why is that?)
• Solving monotone inclusions via compositions of nonexpansive averaged operators
• Signal Recovery by Proximal Forward-Backward Splitting
• A first-order primal-dual algorithm for convex problems with applications to imaging
• Clarke Subgradients of Stratifiable Functions
• Strongly Regular Generalized Equations
• Necessary and sufficient optimality conditions for mathematical programs with equilibrium constraints
• Error bounds in mathematical programming
• On the maximal monotonicity of subdifferential mappings
• An inertial forward-backward-forward primal-dual splitting algorithm for solving monotone inclusion problems
• Inertial Douglas-Rachford splitting for monotone inclusion problems
• A primal-dual splitting method for convex optimization involving Lipschitzian, proximable and linear composite terms
• Generalized equations and their solutions, Part I: Basic theory
• Tangent Cones, Generalized Gradients and Mathematical Programming in Banach Spaces
• Proto-differentiability of set-valued mappings and its applications in optimization
• Sensitivity Analysis of Solutions to Generalized Equations
• A forward-backward splitting method for monotone inclusions without cocoercivity
• The Strong Second-Order Sufficient Condition and Constraint Nondegeneracy in Nonlinear Semidefinite Programming and Their Implications
• Finding best approximation pairs relative to two closed convex sets in Hilbert spaces
• Clarke generalized Jacobian of the projection onto the cone of positive semidefinite matrices
• Sensitivity analysis of generalized equations
• Fifty years of maximal monotonicity
• On the ergodic convergence rates of a first-order primal-dual algorithm
• Lagrangian Duality and Related Multiplier Methods for Variational Inequality Problems
• Title not available (Why is that?)
• A stochastic Bregman primal-dual splitting algorithm for composite optimization
• Generalized Hessian Properties of Regularized Nonsmooth Functions
• From error bounds to the complexity of first-order descent methods for convex functions
• Sensitivity analysis for nonsmooth generalized equations
• Title not available (Why is that?)
• Automatic differentiation of iterative processes
• Sensitivity analysis for mirror-stratifiable convex functions
• The degrees of freedom of partly smooth regularizers
• Bilevel optimization with nonsmooth lower level problems
• Local linear convergence analysis of primal-dual splitting methods
• Convergence of the forward-backward algorithm: beyond the worst-case with the help of geometry
• Sparsifying Transform Learning With Efficient Optimal Updates and Convergence Guarantees
• Conservative set valued fields, automatic differentiation, stochastic gradient methods and deep learning
• Perturbation techniques for convergence analysis of proximal gradient method and other first-order algorithms via variational analysis
• Fixed Point Strategies in Data Science
• Learning consistent discretizations of the total variation
• Learning maximally monotone operators for image recovery
• A primal-dual algorithm with line search for general convex-concave saddle point problems
• Accelerated Bregman Primal-Dual Methods Applied to Optimal Transport and Wasserstein Barycenter Problems
• Sensitivity analysis of maximally monotone inclusions via the proto-differentiability of the resolvent operator
• Convergence of a piggyback-style method for the differentiation of solutions of standard saddle-point problems
• LASSO Reloaded: A Variational Analysis Perspective with Applications to Compressed Sensing

Cited In (1)

• Square root LASSO: well-posedness, Lipschitz stability, and the tuning trade-off

This page was built for publication: Differentiating Nonsmooth Solutions to Parametric Monotone Inclusion Problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6136656)