Semi-proximal Mirror-Prox for Nonsmooth Composite Minimization

Niao He 1 Zaid Harchaoui 2, 3
3 LEAR - Learning and recognition in vision
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, INPG - Institut National Polytechnique de Grenoble
Abstract : We propose a new first-order optimisation algorithm to solve high-dimensional non-smooth composite minimisation problems. Typical examples of such problems have an objective that decomposes into a non-smooth empirical risk part and a non-smooth regularisation penalty. The proposed algorithm, called Semi-Proximal Mirror-Prox, leverages the Fenchel-type representation of one part of the objective while handling the other part of the objective via linear minimization over the domain. The algorithm stands in contrast with more classical proximal gradient algorithms with smoothing, which require the computation of proximal operators at each iteration and can therefore be impractical for high-dimensional problems. We establish the theoretical convergence rate of Semi-Proximal Mirror-Prox, which exhibits the optimal complexity bounds, i.e.O(1/∊2), for the number of calls to linear minimization oracle. We present promising experimental results showing the interest of the approach in comparison to competing methods.
Liste complète des métadonnées

Cited literature [26 references]  Display  Hide  Download

https://hal.inria.fr/hal-01171567
Contributor : Thoth Team <>
Submitted on : Sunday, July 5, 2015 - 4:28:02 PM
Last modification on : Monday, December 17, 2018 - 11:22:02 AM
Document(s) archivé(s) le : Tuesday, April 25, 2017 - 11:34:52 PM

File

main-rpt.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01171567, version 1

Citation

Niao He, Zaid Harchaoui. Semi-proximal Mirror-Prox for Nonsmooth Composite Minimization. NIPS - Advances in Neural Information Processing Systems, Dec 2015, Montreal, Canada. pp.3411-3419. ⟨hal-01171567⟩

Share

Metrics

Record views

944

Files downloads

333