Deep Convolutional Transform Learning - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Deep Convolutional Transform Learning

Résumé

This work introduces a new unsupervised representation learning technique called Deep Convolutional Transform Learning (DCTL). By stacking convolutional transforms, our approach is able to learn a set of independent kernels at different layers. The features extracted in an unsupervised manner can then be used to perform machine learning tasks, such as classification and clustering. The learning technique relies on a well-sounded alternating proximal minimization scheme with established convergence guarantees. Our experimental results show that the proposed DCTL technique outperforms its shallow version CTL, on several benchmark datasets.
Fichier principal
Vignette du fichier
DCTL_iconip_2020_final.pdf (257.28 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02943652 , version 1 (20-09-2020)

Identifiants

  • HAL Id : hal-02943652 , version 1

Citer

Jyoti Maggu, Angshul Majumdar, Emilie Chouzenoux, Giovanni Chierchia. Deep Convolutional Transform Learning. ICONIP 2020 - 27th International Conference on Neural Information Processing, Nov 2020, Bangkok, Thailand. ⟨hal-02943652⟩
98 Consultations
1075 Téléchargements

Partager

Gmail Facebook X LinkedIn More