Structured sparsity-inducing norms through submodular functions - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2010

Structured sparsity-inducing norms through submodular functions

Résumé

Sparse methods for supervised learning aim at finding good linear predictors from as few variables as possible, i.e., with small cardinality of their supports. This combinatorial selection problem is often turned into a convex optimization problem by replacing the cardinality function by its convex envelope (tightest convex lower bound), in this case the L1-norm. In this paper, we investigate more general set-functions than the cardinality, that may incorporate prior knowledge or structural constraints which are common in many applications: namely, we show that for nondecreasing submodular set-functions, the corresponding convex envelope can be obtained from its \lova extension, a common tool in submodular analysis. This defines a family of polyhedral norms, for which we provide generic algorithmic tools (subgradients and proximal operators) and theoretical results (conditions for support recovery or high-dimensional inference). By selecting specific submodular functions, we can give a new interpretation to known norms, such as those based on rank-statistics or grouped norms with potentially overlapping groups; we also define new norms, in particular ones that can be used as non-factorial priors for supervised learning.
Fichier principal
Vignette du fichier
submodular_hal.pdf (586.43 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-00511310 , version 1 (24-08-2010)
hal-00511310 , version 2 (22-09-2010)
hal-00511310 , version 3 (12-11-2010)

Identifiants

Citer

Francis Bach. Structured sparsity-inducing norms through submodular functions. NIPS 2010 : Twenty-Fourth Annual Conference on Neural Information Processing Systems, 2010, Vancouver, Canada. pp.NIPS. ⟨hal-00511310v3⟩
294 Consultations
168 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More