Langevin algorithms for very deep Neural Networks with application to image classification - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2022

Langevin algorithms for very deep Neural Networks with application to image classification

Résumé

Training a very deep neural network is a challenging task, as the deeper a neural network is, the more non-linear it is. We compare the performances of various preconditioned Langevin algorithms with their non-Langevin counterparts for the training of neural networks of increasing depth. For shallow neural networks, Langevin algorithms do not lead to any improvement, however the deeper the network is and the greater are the gains provided by Langevin algorithms. Adding noise to the gradient descent allows to escape from local traps, which are more frequent for very deep neural networks. Following this heuristic we introduce a new Langevin algorithm called Layer Langevin, which consists in adding Langevin noise only to the weights associated to the deepest layers. We then prove the benefits of Langevin and Layer Langevin algorithms for the training of popular deep residual architectures for image classification.
Fichier principal
Vignette du fichier
2212.14718.pdf (244.48 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03980622 , version 1 (09-02-2023)

Licence

Paternité

Identifiants

  • HAL Id : hal-03980622 , version 1

Citer

Pierre Bras. Langevin algorithms for very deep Neural Networks with application to image classification. 2022. ⟨hal-03980622⟩
1 Consultations
4 Téléchargements

Partager

Gmail Facebook X LinkedIn More