TY - GEN
T1 - Graph Classification with 2D Convolutional Neural Networks
AU - Tixier, Antoine J.P.
AU - Nikolentzos, Giannis
AU - Meladianos, Polykarpos
AU - Vazirgiannis, Michalis
N1 - Publisher Copyright:
© Springer Nature Switzerland AG 2019.
PY - 2019/1/1
Y1 - 2019/1/1
N2 - Graph learning is currently dominated by graph kernels, which, while powerful, suffer some significant limitations. Convolutional Neural Networks (CNNs) offer a very appealing alternative, but processing graphs with CNNs is not trivial. To address this challenge, many sophisticated extensions of CNNs have recently been introduced. In this paper, we reverse the problem: rather than proposing yet another graph CNN model, we introduce a novel way to represent graphs as multi-channel image-like structures that allows them to be handled by vanilla 2D CNNs. Experiments reveal that our method is more accurate than state-of-the-art graph kernels and graph CNNs on 4 out of 6 real-world datasets (with and without continuous node attributes), and close elsewhere. Our approach is also preferable to graph kernels in terms of time complexity. Code and data are publicly available (https://github.com/Tixierae/graph_2D_CNN).
AB - Graph learning is currently dominated by graph kernels, which, while powerful, suffer some significant limitations. Convolutional Neural Networks (CNNs) offer a very appealing alternative, but processing graphs with CNNs is not trivial. To address this challenge, many sophisticated extensions of CNNs have recently been introduced. In this paper, we reverse the problem: rather than proposing yet another graph CNN model, we introduce a novel way to represent graphs as multi-channel image-like structures that allows them to be handled by vanilla 2D CNNs. Experiments reveal that our method is more accurate than state-of-the-art graph kernels and graph CNNs on 4 out of 6 real-world datasets (with and without continuous node attributes), and close elsewhere. Our approach is also preferable to graph kernels in terms of time complexity. Code and data are publicly available (https://github.com/Tixierae/graph_2D_CNN).
U2 - 10.1007/978-3-030-30493-5_54
DO - 10.1007/978-3-030-30493-5_54
M3 - Conference contribution
AN - SCOPUS:85072976670
SN - 9783030304928
T3 - Lecture Notes in Computer Science
SP - 578
EP - 593
BT - Artificial Neural Networks and Machine Learning – ICANN 2019
A2 - Kurková, Vera
A2 - Tetko, Igor V.
A2 - Karpov, Pavel
A2 - Theis, Fabian
PB - Springer Verlag
T2 - 28th International Conference on Artificial Neural Networks, ICANN 2019
Y2 - 17 September 2019 through 19 September 2019
ER -