Abstract

Network embedding leverages the node proximity manifested to learn a low-dimensional node vector representation for each node in the network. The learned embeddings could advance various learning tasks such as node classification, network clustering, and link prediction. Most, if not all, of the existing works, are overwhelmingly performed in the context of plain and static networks. Nonetheless, in reality, network structure often evolves over time with addition/deletion of links and nodes. Also, a vast majority of real-world networks are associated with a rich set of node attributes, and their attribute values are also naturally changing, with the emerging of new content patterns and the fading of old content patterns.These changing characteristics motivate us to seek an effective embedding representation to capture network and attribute evolving patterns, which is of fundamental importance for learning in a dynamic environment. To our best knowledge, we are the first to tackle this problem with the following two challenges: (1) the inherently correlated network and node attributes could be noisy and incomplete, it necessitates a robust consensus representation to capture their individual properties and correlations; (2) the embedding learning needs to be performed in an online fashion to adapt to the changes accordingly. In this paper, we tackle this problem by proposing a novel dynamic attributed network embedding framework - DANE. In particular, DANE first provides an offine method for a consensus embedding and then leverages matrix perturbation theory to maintain the freshness of the end embedding results in an online manner. We perform extensive experiments on both synthetic and real attributed networks to corroborate the effectiveness and efficiency of the proposed framework.

Original languageEnglish (US)
Title of host publicationCIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management
PublisherAssociation for Computing Machinery
Pages387-396
Number of pages10
VolumePart F131841
ISBN (Electronic)9781450349185
DOIs
StatePublished - Nov 6 2017
Event26th ACM International Conference on Information and Knowledge Management, CIKM 2017 - Singapore, Singapore
Duration: Nov 6 2017Nov 10 2017

Other

Other26th ACM International Conference on Information and Knowledge Management, CIKM 2017
CountrySingapore
CitySingapore
Period11/6/1711/10/17

Fingerprint

Dynamic environment
Node
Leverage
Clustering
Fading
Perturbation
Prediction
Proximity
Experiment
Network structure
Network dynamics

Keywords

  • Attributed networks
  • Dynamic networks
  • Network embedding

ASJC Scopus subject areas

  • Business, Management and Accounting(all)
  • Decision Sciences(all)

Cite this

Li, J., Dani, H., Hu, X., Tang, J., Chang, Y., & Liu, H. (2017). Attributed network embedding for learning in a dynamic environment. In CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management (Vol. Part F131841, pp. 387-396). Association for Computing Machinery. https://doi.org/10.1145/3132847.3132919

Attributed network embedding for learning in a dynamic environment. / Li, Jundong; Dani, Harsh; Hu, Xia; Tang, Jiliang; Chang, Yi; Liu, Huan.

CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management. Vol. Part F131841 Association for Computing Machinery, 2017. p. 387-396.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, J, Dani, H, Hu, X, Tang, J, Chang, Y & Liu, H 2017, Attributed network embedding for learning in a dynamic environment. in CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management. vol. Part F131841, Association for Computing Machinery, pp. 387-396, 26th ACM International Conference on Information and Knowledge Management, CIKM 2017, Singapore, Singapore, 11/6/17. https://doi.org/10.1145/3132847.3132919
Li J, Dani H, Hu X, Tang J, Chang Y, Liu H. Attributed network embedding for learning in a dynamic environment. In CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management. Vol. Part F131841. Association for Computing Machinery. 2017. p. 387-396 https://doi.org/10.1145/3132847.3132919
Li, Jundong ; Dani, Harsh ; Hu, Xia ; Tang, Jiliang ; Chang, Yi ; Liu, Huan. / Attributed network embedding for learning in a dynamic environment. CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management. Vol. Part F131841 Association for Computing Machinery, 2017. pp. 387-396
@inproceedings{5618101de7894cf68fc0a761543fe318,
title = "Attributed network embedding for learning in a dynamic environment",
abstract = "Network embedding leverages the node proximity manifested to learn a low-dimensional node vector representation for each node in the network. The learned embeddings could advance various learning tasks such as node classification, network clustering, and link prediction. Most, if not all, of the existing works, are overwhelmingly performed in the context of plain and static networks. Nonetheless, in reality, network structure often evolves over time with addition/deletion of links and nodes. Also, a vast majority of real-world networks are associated with a rich set of node attributes, and their attribute values are also naturally changing, with the emerging of new content patterns and the fading of old content patterns.These changing characteristics motivate us to seek an effective embedding representation to capture network and attribute evolving patterns, which is of fundamental importance for learning in a dynamic environment. To our best knowledge, we are the first to tackle this problem with the following two challenges: (1) the inherently correlated network and node attributes could be noisy and incomplete, it necessitates a robust consensus representation to capture their individual properties and correlations; (2) the embedding learning needs to be performed in an online fashion to adapt to the changes accordingly. In this paper, we tackle this problem by proposing a novel dynamic attributed network embedding framework - DANE. In particular, DANE first provides an offine method for a consensus embedding and then leverages matrix perturbation theory to maintain the freshness of the end embedding results in an online manner. We perform extensive experiments on both synthetic and real attributed networks to corroborate the effectiveness and efficiency of the proposed framework.",
keywords = "Attributed networks, Dynamic networks, Network embedding",
author = "Jundong Li and Harsh Dani and Xia Hu and Jiliang Tang and Yi Chang and Huan Liu",
year = "2017",
month = "11",
day = "6",
doi = "10.1145/3132847.3132919",
language = "English (US)",
volume = "Part F131841",
pages = "387--396",
booktitle = "CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Attributed network embedding for learning in a dynamic environment

AU - Li, Jundong

AU - Dani, Harsh

AU - Hu, Xia

AU - Tang, Jiliang

AU - Chang, Yi

AU - Liu, Huan

PY - 2017/11/6

Y1 - 2017/11/6

N2 - Network embedding leverages the node proximity manifested to learn a low-dimensional node vector representation for each node in the network. The learned embeddings could advance various learning tasks such as node classification, network clustering, and link prediction. Most, if not all, of the existing works, are overwhelmingly performed in the context of plain and static networks. Nonetheless, in reality, network structure often evolves over time with addition/deletion of links and nodes. Also, a vast majority of real-world networks are associated with a rich set of node attributes, and their attribute values are also naturally changing, with the emerging of new content patterns and the fading of old content patterns.These changing characteristics motivate us to seek an effective embedding representation to capture network and attribute evolving patterns, which is of fundamental importance for learning in a dynamic environment. To our best knowledge, we are the first to tackle this problem with the following two challenges: (1) the inherently correlated network and node attributes could be noisy and incomplete, it necessitates a robust consensus representation to capture their individual properties and correlations; (2) the embedding learning needs to be performed in an online fashion to adapt to the changes accordingly. In this paper, we tackle this problem by proposing a novel dynamic attributed network embedding framework - DANE. In particular, DANE first provides an offine method for a consensus embedding and then leverages matrix perturbation theory to maintain the freshness of the end embedding results in an online manner. We perform extensive experiments on both synthetic and real attributed networks to corroborate the effectiveness and efficiency of the proposed framework.

AB - Network embedding leverages the node proximity manifested to learn a low-dimensional node vector representation for each node in the network. The learned embeddings could advance various learning tasks such as node classification, network clustering, and link prediction. Most, if not all, of the existing works, are overwhelmingly performed in the context of plain and static networks. Nonetheless, in reality, network structure often evolves over time with addition/deletion of links and nodes. Also, a vast majority of real-world networks are associated with a rich set of node attributes, and their attribute values are also naturally changing, with the emerging of new content patterns and the fading of old content patterns.These changing characteristics motivate us to seek an effective embedding representation to capture network and attribute evolving patterns, which is of fundamental importance for learning in a dynamic environment. To our best knowledge, we are the first to tackle this problem with the following two challenges: (1) the inherently correlated network and node attributes could be noisy and incomplete, it necessitates a robust consensus representation to capture their individual properties and correlations; (2) the embedding learning needs to be performed in an online fashion to adapt to the changes accordingly. In this paper, we tackle this problem by proposing a novel dynamic attributed network embedding framework - DANE. In particular, DANE first provides an offine method for a consensus embedding and then leverages matrix perturbation theory to maintain the freshness of the end embedding results in an online manner. We perform extensive experiments on both synthetic and real attributed networks to corroborate the effectiveness and efficiency of the proposed framework.

KW - Attributed networks

KW - Dynamic networks

KW - Network embedding

UR - http://www.scopus.com/inward/record.url?scp=85037356281&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85037356281&partnerID=8YFLogxK

U2 - 10.1145/3132847.3132919

DO - 10.1145/3132847.3132919

M3 - Conference contribution

VL - Part F131841

SP - 387

EP - 396

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

PB - Association for Computing Machinery

ER -