Skip to Main content Skip to Navigation
Conference papers

Convolutional Kernel Networks

Julien Mairal 1, * Piotr Koniusz 1, * Zaid Harchaoui 1, * Cordelia Schmid 1, *
* Corresponding author
1 LEAR - Learning and recognition in vision
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann, Grenoble INP - Institut polytechnique de Grenoble - Grenoble Institute of Technology
Abstract : An important goal in visual recognition is to devise image representations that are invariant to particular transformations. In this paper, we address this goal with a new type of convolutional neural network (CNN) whose invariance is encoded by a reproducing kernel. Unlike traditional approaches where neural networks are learned either to represent data or for solving a classification task, our network learns to approximate the kernel feature map on training data. Such an approach enjoys several benefits over classical ones. First, by teaching CNNs to be invariant, we obtain simple network architectures that achieve a similar accuracy to more complex ones, while being easy to train and robust to overfitting. Second, we bridge a gap between the neural network literature and kernels, which are natural tools to model invariance. We evaluate our methodology on visual recognition tasks where CNNs have proven to perform well, e.g., digit recognition with the MNIST dataset, and the more challenging CIFAR-10 and STL-10 datasets, where our accuracy is competitive with the state of the art.
Complete list of metadata

Cited literature [31 references]  Display  Hide  Download
Contributor : Julien Mairal Connect in order to contact the contributor
Submitted on : Wednesday, November 12, 2014 - 9:43:50 AM
Last modification on : Thursday, January 20, 2022 - 5:30:14 PM
Long-term archiving on: : Friday, February 13, 2015 - 10:26:01 AM


Files produced by the author(s)


  • HAL Id : hal-01005489, version 3
  • ARXIV : 1406.3332



Julien Mairal, Piotr Koniusz, Zaid Harchaoui, Cordelia Schmid. Convolutional Kernel Networks. Advances in Neural Information Processing Systems (NIPS), Dec 2014, Montreal, Canada. ⟨hal-01005489v3⟩



Les métriques sont temporairement indisponibles