Accelerating inexact successive quadratic approximation for regularized optimization through manifold identification
Author:
Open Access
- 12 January 2023
- journal article
- research article
- Published by Springer Nature in Mathematical Programming
- Vol. 201 (1-2) , 599-633
Abstract
For regularized optimization that minimizes the sum of a smooth term and a regularizer that promotes structured solutions, inexact proximal-Newton-type methods, or successive quadratic approximation (SQA) methods, are widely used for their superlinear convergence in terms of iterations. However, unlike the counter parts in smooth optimization, they suffer from lengthy running time in solving regularized subproblems because even approximate solutions cannot be computed easily, so their empirical time cost is not as impressive. In this work, we first show that for partly smooth regularizers, although general inexact solutions cannot identify the active manifold that makes the objective function smooth, approximate solutions generated by commonly-used subproblem solvers will identify this manifold, even with arbitrarily low solution precision. We then utilize this property to propose an improved SQA method, $$^{+}$$ , that switches to efficient smooth optimization methods after this manifold is identified. We show that for a wide class of degenerate solutions, $$^{+}$$ possesses superlinear convergence not only in iterations, but also in running time because the cost per iteration is bounded. In particular, our superlinear convergence result holds on problems satisfying a sharpness condition that is more general than that in existing literature. We also prove iterate convergence under a sharpness condition for inexact SQA, which is novel for this family of methods that could easily violate the classical relative-error condition frequently used in proving convergence under similar conditions. Experiments on real-world problems support that $$^{+}$$ improves running time over some modern solvers for regularized optimization.
Keywords
All Related Versions
Funding Information
- National Science and Technology Council (109-2222-E-001-003 -MY3, 111-2628-E-001-003-)
- Academia Sinica (AS-GCS- 111-M05)
This publication has 42 references indexed in Scilit:
- Accelerated Block-coordinate Relaxation for Regularized OptimizationSIAM Journal on Optimization, 2012
- Exact Matrix Completion via Convex OptimizationFoundations of Computational Mathematics, 2009
- A coordinate gradient descent method for nonsmooth separable minimizationMathematical Programming, 2007
- Stable signal recovery from incomplete and inaccurate measurementsCommunications on Pure and Applied Mathematics, 2006
- Functions and Sets of Smooth Substructure: Relationships and ExamplesComputational Optimization and Applications, 2005
- Newton methods for nonsmooth convex minimization: connections among -Lagrangian, Riemannian Newton and SQP methodsMathematical Programming, 2005
- A -algorithm for convex minimizationMathematical Programming, 2005
- Active Sets, Nonsmoothness, and SensitivitySIAM Journal on Optimization, 2002
- On $\mathcalVU$-theory for Functions with Primal-Dual Gradient StructureSIAM Journal on Optimization, 2000
- On gradients of functions definable in o-minimal structuresAnnales de l'institut Fourier, 1998