MOTIFNET: A MOTIF-BASED GRAPH CONVOLUTIONAL NETWORK for DIRECTED GRAPHS

Federico Monti, Karl Otness, Michael M. Bronstein

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

102 Scopus citations

Abstract

Deep learning on graphs and in particular, graph convolutional neural networks, have recently attracted significant attention in the machine learning community. Many of such techniques explore the analogy between the graph Laplacian eigenvectors and the classical Fourier basis, allowing to formulate the convolution as a multiplication in the spectral domain. One of the key drawback of spectral CNNs is their explicit assumption of an undirected graph, leading to a symmetric Laplacian matrix with orthogonal eigendecomposition. In this work we propose MotifNet, a graph CNN capable of dealing with directed graphs by exploiting local graph motifs. We present experimental evidence showing the advantage of our approach on real data.

Original languageEnglish
Title of host publication2018 IEEE Data Science Workshop, DSW 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages225-228
Number of pages4
ISBN (Print)9781538644102
DOIs
StatePublished - 17 Aug 2018
Event2018 IEEE Data Science Workshop, DSW 2018 - Lausanne, Switzerland
Duration: 4 Jun 20186 Jun 2018

Publication series

Name2018 IEEE Data Science Workshop, DSW 2018 - Proceedings

Conference

Conference2018 IEEE Data Science Workshop, DSW 2018
Country/TerritorySwitzerland
CityLausanne
Period4/06/186/06/18

Funding

FundersFunder number
Horizon 2020 Framework Programme724228

    Keywords

    • Directed Graphs
    • Geometric Deep Learning
    • Graph Convolutional Neural Networks
    • Graph Motifs

    Fingerprint

    Dive into the research topics of 'MOTIFNET: A MOTIF-BASED GRAPH CONVOLUTIONAL NETWORK for DIRECTED GRAPHS'. Together they form a unique fingerprint.

    Cite this