End-to-End Incremental Learning

Abstract : Although deep learning approaches have stood out in recent years due to their state-of-the-art results, they continue to suffer from catastrophic forgetting, a dramatic decrease in overall performance when training with new classes added incrementally. This is due to current neural network architectures requiring the entire dataset, consisting of all the samples from the old as well as the new classes, to update the model---a requirement that becomes easily unsustainable as the number of classes grows. We address this issue with our approach to learn deep neural networks incrementally, using new data and only a small exemplar set corresponding to samples from the old classes. This is based on a loss composed of a distillation measure to retain the knowledge acquired from the old classes, and a cross-entropy loss to learn the new classes. Our incremental training is achieved while keeping the entire framework end-to-end, i.e., learning the data representation and the classifier jointly, unlike recent methods with no such guarantees. We evaluate our method extensively on the CIFAR-100 and ImageNet (ILSVRC 2012) image classification datasets, and show state-of-the-art performance.
Type de document :
Communication dans un congrès
Vittorio Ferrari; Martial Hebert; Cristian Sminchisescu; Yair Weiss. ECCV - European Conference on Computer Vision, Sep 2018, Munich, Germany. 2018, Lecture Notes in Computer Science
Liste complète des métadonnées

https://hal.inria.fr/hal-01849366
Contributeur : Karteek Alahari <>
Soumis le : jeudi 26 juillet 2018 - 10:05:27
Dernière modification le : dimanche 5 août 2018 - 21:34:09

Fichier

IncrementalLearning_ECCV2018.p...
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01849366, version 1
  • ARXIV : 1807.09536

Collections

Citation

Francisco Castro, Manuel Marín-Jiménez, Nicolás Guil, Cordelia Schmid, Karteek Alahari. End-to-End Incremental Learning. Vittorio Ferrari; Martial Hebert; Cristian Sminchisescu; Yair Weiss. ECCV - European Conference on Computer Vision, Sep 2018, Munich, Germany. 2018, Lecture Notes in Computer Science. 〈hal-01849366〉

Partager

Métriques

Consultations de la notice

155

Téléchargements de fichiers

59