On the stable recovery of deep structured linear networks under sparsity constraints - IRT Saint Exupéry - Institut de Recherche Technologique Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

On the stable recovery of deep structured linear networks under sparsity constraints

Résumé

We consider a deep structured linear network under sparsity constraints. We study sharp conditions guaranteeing the stability of the optimal parameters defining the network. More precisely, we provide sharp conditions on the network architecture and the sample under which the error on the parameters defining the network scales linearly with the reconstruction error (i.e. the risk). Therefore, under these conditions, the weights obtained with a successful algorithms are well defined and only depend on the architecture of the network and the sample. The features in the latent spaces are stably defined. The stability property is required in order to interpret the features defined in the latent spaces. It can also lead to a guarantee on the statistical risk. This is what motivates this study. The analysis is based on the recently proposed Tensorial Lifting. The particularity of this paper is to consider a sparsity prior. This leads to a better stability constant. As an illustration, we detail the analysis and provide sharp stability guarantees for convolutional linear network under sparsity prior. In this analysis, we distinguish the role of the network architecture and the sample input. This highlights the requirements on the data in connection to parameter stability.
Fichier principal
Vignette du fichier
sparse_net.pdf (256.45 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01526083 , version 1 (31-05-2017)
hal-01526083 , version 2 (19-02-2018)
hal-01526083 , version 3 (15-05-2020)

Identifiants

Citer

François Malgouyres. On the stable recovery of deep structured linear networks under sparsity constraints. Mathematical and Scientific Machine Learning, Jul 2020, Princeton, United States. ⟨hal-01526083v3⟩
341 Consultations
406 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More