Invariance and Stability of Deep Convolutional Representations

Abstract : In this paper, we study deep signal representations that are near-invariant to groups of transformations and stable to the action of diffeomorphisms without losing signal information. This is achieved by generalizing the multilayer kernel introduced in the context of convolutional kernel networks and by studying the geometry of the corresponding reproducing kernel Hilbert space. We show that the signal representation is stable, and that models from this functional space, such as a large class of convolutional neural networks, may enjoy the same stability.
Liste complète des métadonnées

Cited literature [33 references]  Display  Hide  Download

https://hal.inria.fr/hal-01630265
Contributor : Alberto Bietti <>
Submitted on : Tuesday, November 7, 2017 - 1:53:07 PM
Last modification on : Thursday, February 7, 2019 - 4:31:29 PM
Document(s) archivé(s) le : Thursday, February 8, 2018 - 1:38:57 PM

Identifiers

  • HAL Id : hal-01630265, version 1

Collections

Citation

Alberto Bietti, Julien Mairal. Invariance and Stability of Deep Convolutional Representations. NIPS 2017 - 31st Conference on Advances in Neural Information Processing Systems, Dec 2017, Los Angeles, CA, United States. pp.1622-1632. ⟨hal-01630265⟩

Share

Metrics

Record views

739

Files downloads

636