Riemannian proximal gradient methods
From MaRDI portal
Publication:2149554
DOI10.1007/S10107-021-01632-3zbMATH Open1492.90012arXiv1909.06065OpenAlexW3135904811WikidataQ115385307 ScholiaQ115385307MaRDI QIDQ2149554FDOQ2149554
Publication date: 29 June 2022
Published in: Mathematical Programming. Series A. Series B (Search for Journal in Brave)
Abstract: In the Euclidean setting, the proximal gradient method and its accelerated variants are a class of efficient algorithms for optimization problems with decomposable objective. In this paper, we develop a Riemannian proximal gradient method (RPG) and its accelerated variant (ARPG) for similar problems but constrained on a manifold. The global convergence of RPG has been established under mild assumptions, and the O(1/k) is also derived for RPG based on the notion of retraction convexity. If assuming the objective function obeys the Rimannian Kurdyka-Lojasiewicz (KL) property, it is further shown that the sequence generated by RPG converges to a single stationary point. As in the Euclidean setting, local convergence rate can be established if the objective function satisfies the Riemannian KL property with an exponent. Moreover, we have shown that the restriction of a semialgebraic function onto the Stiefel manifold satisfies the Riemannian KL property, which covers for example the well-known sparse PCA problem. Numerical experiments on random and synthetic data are conducted to test the performance of the proposed RPG and ARPG.
Full work available at URL: https://arxiv.org/abs/1909.06065
Computational methods for problems pertaining to statistics (62-08) Nonconvex programming, global optimization (90C26) Computational methods for problems pertaining to operations research and mathematical programming (90-08)
Cites Work
- Title not available (Why is that?)
- Title not available (Why is that?)
- Title not available (Why is that?)
- Title not available (Why is that?)
- Title not available (Why is that?)
- Title not available (Why is that?)
- A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems
- Functional and shape data analysis
- First-Order Methods in Optimization
- Proximal Alternating Minimization and Projection Methods for Nonconvex Problems: An Approach Based on the Kurdyka-Łojasiewicz Inequality
- Proximal alternating linearized minimization for nonconvex and nonsmooth problems
- Clarke Subgradients of Stratifiable Functions
- Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward-backward splitting, and regularized Gauss-Seidel methods
- Accelerated gradient methods for nonconvex nonlinear and stochastic programming
- \(\varepsilon\)-subgradient algorithms for locally Lipschitz functions on Riemannian manifolds
- Proximal Point Algorithm On Riemannian Manifolds
- The Łojasiewicz Inequality for Nonsmooth Subanalytic Functions with Applications to Subgradient Dynamical Systems
- Generalized gradients and characterization of epi-Lipschitz sets in Riemannian manifolds
- Proof of the gradient conjecture of R. Thom.
- A Broyden Class of Quasi-Newton Methods for Riemannian Optimization
- Pointwise convergence of gradient‐like systems
- A splitting method for orthogonality constrained problems
- An Extrinsic Look at the Riemannian Hessian
- Iteration-complexity of gradient, subgradient and proximal point methods on Riemannian manifolds
- Simple algorithms for optimization on Riemannian manifolds with constraints
- Introduction to Riemannian Manifolds
- Global rates of convergence for nonconvex optimization on manifolds
- Nonsmooth trust region algorithms for locally Lipschitz functions on Riemannian manifolds
- A Riemannian Gradient Sampling Algorithm for Nonsmooth Optimization on Manifolds
- An Augmented Lagrangian Method for $\ell_{1}$-Regularized Optimization Problems with Orthogonality Constraints
- Self-contracted curves in Riemannian manifolds
- Line Search Algorithms for Locally Lipschitz Functions on Riemannian Manifolds
- Proximal Gradient Method for Nonsmooth Optimization over the Stiefel Manifold
- Computing Riemannian center of mass on Hadamard manifolds
- Weakly Correlated Sparse Components with Nearly Orthonormal Loadings
Cited In (23)
- An inexact Riemannian proximal gradient method
- Riemannian gradient algorithm for the numerical solution of Stein equations
- An image inpainting algorithm using exemplar matching and low-rank sparse prior
- A communication-efficient and privacy-aware distributed algorithm for sparse PCA
- Smoothing algorithms for nonsmooth optimization over the Stiefel manifold with applications to the graph Fourier basis problem
- Proximal quasi-Newton method for composite optimization over the Stiefel manifold
- Weakly Convex Optimization over Stiefel Manifold Using Riemannian Subgradient-Type Methods
- A new Bayesian approach to global optimization on parametrized surfaces in \(\mathbb{R}^3\)
- Proximal point method for quasiconvex functions in Riemannian manifolds
- Zeroth-order Riemannian averaging stochastic approximation algorithms
- An adaptive regularized proximal Newton-type methods for composite optimization over the Stiefel manifold
- Riemannian thresholding methods for row-sparse and low-rank matrix recovery
- Stochastic Gradient Descent on Riemannian Manifolds
- A Riemannian Proximal Newton Method
- Proximal linear maps
- Proximal gradient method for nonconvex and nonsmooth optimization on Hadamard manifolds
- A Dynamic Smoothing Technique for a Class of Nonsmooth Optimization Problems on Manifolds
- Nonsmooth optimization over the Stiefel manifold and beyond: proximal gradient method and recent variants
- Practical gradient and conjugate gradient methods on flag manifolds
- A penalty-free infeasible approach for a class of nonsmooth optimization problems over the Stiefel manifold
- Proximal gradient algorithm with trust region scheme on Riemannian manifold
- Riemannian Stochastic Variance Reduced Gradient Algorithm with Retraction and Vector Transport
- Proximal Point Algorithm On Riemannian Manifolds
Uses Software
This page was built for publication: Riemannian proximal gradient methods
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2149554)
