MGAE

marginalized graph autoencoder for graph clustering

Chun Wang, Shirui Pan, Guodong Long, Xingquan Zhu, Jing Jiang

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

Abstract

Graph clustering aims to discover community structures in networks, the task being fundamentally challenging mainly because the topology structure and the content of the graphs are dicult to represent for clustering analysis. Recently, graph clustering has moved from traditional shallow methods to deep learning approaches, thanks to the unique feature representation learning capability of deep learning. However, existing deep approaches for graph clustering can only exploit the structure information, while ignoring the content information associated with the nodes in a graph. In this paper, we propose a novel marginalized graph autoencoder (MGAE) algorithm for graph clustering. The key innovation of MGAE is that it advances the autoencoder to the graph domain, so graph representation learning can be carried out not only in a purely unsupervised se.ing by leveraging structure and content information, it can also be stacked in a deep fashion to learn effective representation. From a technical viewpoint, we propose a marginalized graph convolutional network to corrupt network node content, allowing node content to interact with network features, and marginalizes the corrupted features in a graph autoencoder context to learn graph feature representations. The learned features are fed into the spectral clustering algorithm for graph clustering. Experimental results on benchmark datasets demonstrate the superior performance of MGAE, compared to numerous baselines.

Original languageEnglish
Title of host publicationCIKM'17 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management
Subtitle of host publicationNovember 6–10, 2017 Singapore, Singapore
EditorsMark Sanderson, Ada Fu, Jimeng Sun
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Pages889-898
Number of pages10
ISBN (Electronic)9781450349185
DOIs
Publication statusPublished - 2017
Externally publishedYes
EventACM International Conference on Information and Knowledge Management 2017 - Singapore, Singapore
Duration: 6 Nov 201710 Nov 2017
Conference number: 26th
http://www.cikmconference.org/CIKM2017/

Conference

ConferenceACM International Conference on Information and Knowledge Management 2017
Abbreviated titleCIKM 2017
CountrySingapore
CitySingapore
Period6/11/1710/11/17
Internet address

Keywords

  • Autoencoder
  • Graph autoencoder
  • Graph clustering
  • Graph convolutional network
  • Network representation

Cite this

Wang, C., Pan, S., Long, G., Zhu, X., & Jiang, J. (2017). MGAE: marginalized graph autoencoder for graph clustering. In M. Sanderson, A. Fu, & J. Sun (Eds.), CIKM'17 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management: November 6–10, 2017 Singapore, Singapore (pp. 889-898). New York NY USA: Association for Computing Machinery (ACM). https://doi.org/10.1145/3132847.3132967
Wang, Chun ; Pan, Shirui ; Long, Guodong ; Zhu, Xingquan ; Jiang, Jing. / MGAE : marginalized graph autoencoder for graph clustering. CIKM'17 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management: November 6–10, 2017 Singapore, Singapore. editor / Mark Sanderson ; Ada Fu ; Jimeng Sun. New York NY USA : Association for Computing Machinery (ACM), 2017. pp. 889-898
@inproceedings{2110eb02c20447e0b13da6f62c7ae963,
title = "MGAE: marginalized graph autoencoder for graph clustering",
abstract = "Graph clustering aims to discover community structures in networks, the task being fundamentally challenging mainly because the topology structure and the content of the graphs are dicult to represent for clustering analysis. Recently, graph clustering has moved from traditional shallow methods to deep learning approaches, thanks to the unique feature representation learning capability of deep learning. However, existing deep approaches for graph clustering can only exploit the structure information, while ignoring the content information associated with the nodes in a graph. In this paper, we propose a novel marginalized graph autoencoder (MGAE) algorithm for graph clustering. The key innovation of MGAE is that it advances the autoencoder to the graph domain, so graph representation learning can be carried out not only in a purely unsupervised se.ing by leveraging structure and content information, it can also be stacked in a deep fashion to learn effective representation. From a technical viewpoint, we propose a marginalized graph convolutional network to corrupt network node content, allowing node content to interact with network features, and marginalizes the corrupted features in a graph autoencoder context to learn graph feature representations. The learned features are fed into the spectral clustering algorithm for graph clustering. Experimental results on benchmark datasets demonstrate the superior performance of MGAE, compared to numerous baselines.",
keywords = "Autoencoder, Graph autoencoder, Graph clustering, Graph convolutional network, Network representation",
author = "Chun Wang and Shirui Pan and Guodong Long and Xingquan Zhu and Jing Jiang",
year = "2017",
doi = "10.1145/3132847.3132967",
language = "English",
pages = "889--898",
editor = "Sanderson, {Mark } and Ada Fu and Jimeng Sun",
booktitle = "CIKM'17 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management",
publisher = "Association for Computing Machinery (ACM)",
address = "United States of America",

}

Wang, C, Pan, S, Long, G, Zhu, X & Jiang, J 2017, MGAE: marginalized graph autoencoder for graph clustering. in M Sanderson, A Fu & J Sun (eds), CIKM'17 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management: November 6–10, 2017 Singapore, Singapore. Association for Computing Machinery (ACM), New York NY USA, pp. 889-898, ACM International Conference on Information and Knowledge Management 2017, Singapore, Singapore, 6/11/17. https://doi.org/10.1145/3132847.3132967

MGAE : marginalized graph autoencoder for graph clustering. / Wang, Chun; Pan, Shirui; Long, Guodong; Zhu, Xingquan; Jiang, Jing.

CIKM'17 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management: November 6–10, 2017 Singapore, Singapore. ed. / Mark Sanderson; Ada Fu; Jimeng Sun. New York NY USA : Association for Computing Machinery (ACM), 2017. p. 889-898.

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

TY - GEN

T1 - MGAE

T2 - marginalized graph autoencoder for graph clustering

AU - Wang, Chun

AU - Pan, Shirui

AU - Long, Guodong

AU - Zhu, Xingquan

AU - Jiang, Jing

PY - 2017

Y1 - 2017

N2 - Graph clustering aims to discover community structures in networks, the task being fundamentally challenging mainly because the topology structure and the content of the graphs are dicult to represent for clustering analysis. Recently, graph clustering has moved from traditional shallow methods to deep learning approaches, thanks to the unique feature representation learning capability of deep learning. However, existing deep approaches for graph clustering can only exploit the structure information, while ignoring the content information associated with the nodes in a graph. In this paper, we propose a novel marginalized graph autoencoder (MGAE) algorithm for graph clustering. The key innovation of MGAE is that it advances the autoencoder to the graph domain, so graph representation learning can be carried out not only in a purely unsupervised se.ing by leveraging structure and content information, it can also be stacked in a deep fashion to learn effective representation. From a technical viewpoint, we propose a marginalized graph convolutional network to corrupt network node content, allowing node content to interact with network features, and marginalizes the corrupted features in a graph autoencoder context to learn graph feature representations. The learned features are fed into the spectral clustering algorithm for graph clustering. Experimental results on benchmark datasets demonstrate the superior performance of MGAE, compared to numerous baselines.

AB - Graph clustering aims to discover community structures in networks, the task being fundamentally challenging mainly because the topology structure and the content of the graphs are dicult to represent for clustering analysis. Recently, graph clustering has moved from traditional shallow methods to deep learning approaches, thanks to the unique feature representation learning capability of deep learning. However, existing deep approaches for graph clustering can only exploit the structure information, while ignoring the content information associated with the nodes in a graph. In this paper, we propose a novel marginalized graph autoencoder (MGAE) algorithm for graph clustering. The key innovation of MGAE is that it advances the autoencoder to the graph domain, so graph representation learning can be carried out not only in a purely unsupervised se.ing by leveraging structure and content information, it can also be stacked in a deep fashion to learn effective representation. From a technical viewpoint, we propose a marginalized graph convolutional network to corrupt network node content, allowing node content to interact with network features, and marginalizes the corrupted features in a graph autoencoder context to learn graph feature representations. The learned features are fed into the spectral clustering algorithm for graph clustering. Experimental results on benchmark datasets demonstrate the superior performance of MGAE, compared to numerous baselines.

KW - Autoencoder

KW - Graph autoencoder

KW - Graph clustering

KW - Graph convolutional network

KW - Network representation

UR - http://www.scopus.com/inward/record.url?scp=85037344286&partnerID=8YFLogxK

U2 - 10.1145/3132847.3132967

DO - 10.1145/3132847.3132967

M3 - Conference Paper

SP - 889

EP - 898

BT - CIKM'17 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management

A2 - Sanderson, Mark

A2 - Fu, Ada

A2 - Sun, Jimeng

PB - Association for Computing Machinery (ACM)

CY - New York NY USA

ER -

Wang C, Pan S, Long G, Zhu X, Jiang J. MGAE: marginalized graph autoencoder for graph clustering. In Sanderson M, Fu A, Sun J, editors, CIKM'17 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management: November 6–10, 2017 Singapore, Singapore. New York NY USA: Association for Computing Machinery (ACM). 2017. p. 889-898 https://doi.org/10.1145/3132847.3132967