Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphs

Martin Simonovsky (1, 2, 3) , Nikos Komodakis (1, 2, 3)

1 ENPC - École des Ponts ParisTech
2 imagine [Marne-la-Vallée]
3 LIGM - Laboratoire d'Informatique Gaspard-Monge

A number of problems can be formulated as prediction on graph-structured data. In this work, we generalize the convolution operator from regular grids to arbitrary graphs while avoiding the spectral domain, which allows us to handle graphs of varying size and connec-tivity. To move beyond a simple diffusion, filter weights are conditioned on the specific edge labels in the neighborhood of a vertex. Together with the proper choice of graph coarsening, we explore constructing deep neural networks for graph classification. In particular, we demonstrate the generality of our formulation in point cloud classification , where we set the new state of the art, and on a graph classification dataset, where we outperform other deep learning approaches. The source code is available at https://github.com/mys007/ecc.

Domaines: Informatique [cs]

Martin Simonovsky, Nikos Komodakis. Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphs. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2017), Jul 2017, Honolulu, United States. ⟨hal-01576919⟩