Numerous important problems can be framed as learning from graph data. We
propose a framework for learning convolutional neural networks for arbitrary
graphs. These graphs may be undirected, directed, and with both discrete and
continuous node and edge attributes. Analogous to image-based convolutional
networks that operate on locally connected regions of the input, we present a
general approach to extracting locally connected regions from graphs. Using
established benchmark data sets, we demonstrate that the learned feature
representations are competitive with state of the art graph kernels and that
their computation is highly efficient.
Description
[1605.05273] Learning Convolutional Neural Networks for Graphs
%0 Generic
%1 niepert2016learning
%A Niepert, Mathias
%A Ahmed, Mohamed
%A Kutzkov, Konstantin
%D 2016
%K 2016 cnn deep-learning graph icml
%T Learning Convolutional Neural Networks for Graphs
%U http://arxiv.org/abs/1605.05273
%X Numerous important problems can be framed as learning from graph data. We
propose a framework for learning convolutional neural networks for arbitrary
graphs. These graphs may be undirected, directed, and with both discrete and
continuous node and edge attributes. Analogous to image-based convolutional
networks that operate on locally connected regions of the input, we present a
general approach to extracting locally connected regions from graphs. Using
established benchmark data sets, we demonstrate that the learned feature
representations are competitive with state of the art graph kernels and that
their computation is highly efficient.
@misc{niepert2016learning,
abstract = {Numerous important problems can be framed as learning from graph data. We
propose a framework for learning convolutional neural networks for arbitrary
graphs. These graphs may be undirected, directed, and with both discrete and
continuous node and edge attributes. Analogous to image-based convolutional
networks that operate on locally connected regions of the input, we present a
general approach to extracting locally connected regions from graphs. Using
established benchmark data sets, we demonstrate that the learned feature
representations are competitive with state of the art graph kernels and that
their computation is highly efficient.},
added-at = {2018-07-20T13:11:16.000+0200},
author = {Niepert, Mathias and Ahmed, Mohamed and Kutzkov, Konstantin},
biburl = {https://www.bibsonomy.org/bibtex/206019e4bf9a511bd68027aa1f3f96084/analyst},
description = {[1605.05273] Learning Convolutional Neural Networks for Graphs},
interhash = {551b00a837281b8cdd72b3da25476293},
intrahash = {06019e4bf9a511bd68027aa1f3f96084},
keywords = {2016 cnn deep-learning graph icml},
note = {cite arxiv:1605.05273Comment: To be presented at ICML 2016},
timestamp = {2018-07-20T13:11:16.000+0200},
title = {Learning Convolutional Neural Networks for Graphs},
url = {http://arxiv.org/abs/1605.05273},
year = 2016
}