A smoothing approach for composite conditional gradient with nonsmooth loss - Inria - Institut national de recherche en sciences et technologies du numérique
Reports (Research Report) Year : 2014

A smoothing approach for composite conditional gradient with nonsmooth loss

Abstract

We consider learning problems where the non-smoothness lies both in the convex empirical risk and in the regularization penalty. Examples of such problems include learning with nonsmooth loss functions and atomic decomposition regularization penalty. Such doubly nonsmooth learning problems prevent the use of recently proposed composite conditional gradient algorithms for training, which are particularly attractive for large-scale applications. Indeed, they rely on the assumption that the empirical risk part of the objective is smooth. We propose a composite conditional gradient algorithm with smoothing to tackle such learning problems. We set up a framework allowing to systematically design parametrized smooth surrogates of nonsmooth loss functions. We then propose a smoothed composite conditional gradient algorithm, for which we prove theoretical guarantees on the accuracy. We present promising experimental results on collaborative filtering tasks.
Fichier principal
Vignette du fichier
RR-8662.pdf (881.38 Ko) Télécharger le fichier
Origin Files produced by the author(s)
Loading...

Dates and versions

hal-01096630 , version 1 (15-01-2015)

Identifiers

  • HAL Id : hal-01096630 , version 1

Cite

Federico Pierucci, Zaid Harchaoui, Jérôme Malick. A smoothing approach for composite conditional gradient with nonsmooth loss. [Research Report] RR-8662, INRIA Grenoble. 2014. ⟨hal-01096630⟩
736 View
534 Download

Share

More