Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering

06/30/2016
by   Michaël Defferrard, et al.
0

In this work, we are interested in generalizing convolutional neural networks (CNNs) from low-dimensional regular grids, where image, video and speech are represented, to high-dimensional irregular domains, such as social networks, brain connectomes or words' embedding, represented by graphs. We present a formulation of CNNs in the context of spectral graph theory, which provides the necessary mathematical background and efficient numerical schemes to design fast localized convolutional filters on graphs. Importantly, the proposed technique offers the same linear computational complexity and constant learning complexity as classical CNNs, while being universal to any graph structure. Experiments on MNIST and 20NEWS demonstrate the ability of this novel deep learning system to learn local, stationary, and compositional features on graphs.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
06/03/2016

Generalizing the Convolution Operator to extend CNNs to Irregular Domains

Convolutional Neural Networks (CNNs) have become the state-of-the-art in...
research
12/21/2013

Spectral Networks and Locally Connected Networks on Graphs

Convolutional Neural Networks are extremely efficient architectures in i...
research
09/28/2016

Graph Based Convolutional Neural Network

The benefit of localized features within the regular domain has given ri...
research
01/10/2018

Adaptive Graph Convolutional Neural Networks

Graph Convolutional Neural Networks (Graph CNNs) are generalizations of ...
research
07/29/2017

Classifying Graphs as Images with Convolutional Neural Networks

The task of graph classification is currently dominated by graph kernels...
research
12/02/2016

SyncSpecCNN: Synchronized Spectral CNN for 3D Shape Segmentation

In this paper, we study the problem of semantic annotation on 3D models ...
research
05/13/2017

Deep neural networks on graph signals for brain imaging analysis

Brain imaging data such as EEG or MEG are high-dimensional spatiotempora...

Please sign up or login with your details

Forgot password? Click here to reset