Model Consistency for Learning with Mirror-Stratifiable Regularizers

Abstract : Low-complexity non-smooth convex regular-izers are routinely used to impose some structure (such as sparsity or low-rank) on the coefficients for linear predictors in supervised learning. Model consistency consists then in selecting the correct structure (for instance support or rank) by regularized empirical risk minimization. It is known that model consistency holds under appropriate non-degeneracy conditions. However such conditions typically fail for highly correlated designs and it is observed that regularization methods tend to select larger models. In this work, we provide the theoretical underpinning of this behavior using the notion of mirror-stratifiable regular-izers. This class of regularizers encompasses the most well-known in the literature, including the 1 or trace norms. It brings into play a pair of primal-dual models, which in turn allows one to locate the structure of the solution using a specific dual certificate. We also show how this analysis is applicable to optimal solutions of the learning problem, and also to the iterates computed by a certain class of stochastic proximal-gradient algorithms.
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01988309
Contributeur : Jalal Fadili <>
Soumis le : lundi 21 janvier 2019 - 16:30:38
Dernière modification le : jeudi 7 février 2019 - 15:41:38

Fichier

stratificationAISTATS.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01988309, version 1

Citation

Jalal M. Fadili, Guillaume Garrigos, Jérôme Malick, Gabriel Peyré. Model Consistency for Learning with Mirror-Stratifiable Regularizers. International Conference on Artificial Intelligence and Statistics (AISTATS), Apr 2019, Naha, Japan. ⟨hal-01988309⟩

Partager

Métriques

Consultations de la notice

138

Téléchargements de fichiers

27