Singleshot : a scalable Tucker tensor decomposition - Normandie Université Accéder directement au contenu
Communication Dans Un Congrès Année : 2019

Singleshot : a scalable Tucker tensor decomposition

Résumé

This paper introduces a new approach for the scalable Tucker decomposition problem. Given a tensor X , the algorithm proposed, named Singleshot, allows to perform the inference task by processing one subtensor drawn from X at a time. The key principle of our approach is based on the recursive computations of the gradient and on cyclic update of the latent factors involving only one single step of gradient descent. We further improve the computational efficiency of Singleshot by proposing an inexact gradient version named Singleshotinexact. The two algorithms are backed with theoretical guarantees of convergence and convergence rates under mild conditions. The scalabilty of the proposed approaches, which can be easily extended to handle some common constraints encountered in tensor decomposition (e.g non-negativity), is proven via numerical experiments on both synthetic and real data sets.
Fichier principal
Vignette du fichier
SingleshotTucker (1).pdf (622.73 Ko) Télécharger le fichier
SupplementarymaterialSingleshotTucker.pdf (633.21 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02288245 , version 1 (14-09-2019)
hal-02288245 , version 2 (26-10-2019)

Identifiants

  • HAL Id : hal-02288245 , version 1

Citer

Abraham Traoré, Maxime Berar, Alain Rakotomamonjy. Singleshot : a scalable Tucker tensor decomposition. Neurips-2019, Dec 2019, Vancouver, Canada. ⟨hal-02288245v1⟩
186 Consultations
277 Téléchargements

Partager

Gmail Facebook X LinkedIn More