Network Embedding with Deep Metric Learning

Xiaotao CHENG; Lixin JI; Ruiyang HUANG; Ruifei CUI

doi:10.1587/transinf.2018EDP7233

IEICE TRANSACTIONS on Information

Network Embedding with Deep Metric Learning

Xiaotao CHENG, Lixin JI, Ruiyang HUANG, Ruifei CUI

Full Text Views

0

Cite this

Summary :

Network embedding has attracted an increasing amount of attention in recent years due to its wide-ranging applications in graph mining tasks such as vertex classification, community detection, and network visualization. Network embedding is an important method to learn low-dimensional representations of vertices in networks, aiming to capture and preserve the network structure. Almost all the existing network embedding methods adopt the so-called Skip-gram model in Word2vec. However, as a bag-of-words model, the skip-gram model mainly utilized the local structure information. The lack of information metrics for vertices in global network leads to the mix of vertices with different labels in the new embedding space. To solve this problem, in this paper we propose a Network Representation Learning method with Deep Metric Learning, namely DML-NRL. By setting the initialized anchor vertices and adding the similarity measure in the training progress, the distance information between different labels of vertices in the network is integrated into the vertex representation, which improves the accuracy of network embedding algorithm effectively. We compare our method with baselines by applying them to the tasks of multi-label classification and data visualization of vertices. The experimental results show that our method outperforms the baselines in all three datasets, and the method has proved to be effective and robust.

Publication: IEICE TRANSACTIONS on Information Vol.E102-D No.3 pp.568-578

Publication Date: 2019/03/01

Publicized: 2018/12/26

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2018EDP7233

Type of Manuscript: PAPER

Category: Artificial Intelligence, Data Mining

Authors

Xiaotao CHENG
  National Digital Switching System Engineering & Technological R&D Center
Lixin JI
  National Digital Switching System Engineering & Technological R&D Center
Ruiyang HUANG
  National Digital Switching System Engineering & Technological R&D Center
Ruifei CUI
  Radboud University Nijmegen

Keyword

deep metric learning, network representation learning, likelihood label, anchor initialization, semi-supervised learning

Cite this

Copy

Xiaotao CHENG, Lixin JI, Ruiyang HUANG, Ruifei CUI, "Network Embedding with Deep Metric Learning" in IEICE TRANSACTIONS on Information, vol. E102-D, no. 3, pp. 568-578, March 2019, doi: 10.1587/transinf.2018EDP7233.
Abstract: Network embedding has attracted an increasing amount of attention in recent years due to its wide-ranging applications in graph mining tasks such as vertex classification, community detection, and network visualization. Network embedding is an important method to learn low-dimensional representations of vertices in networks, aiming to capture and preserve the network structure. Almost all the existing network embedding methods adopt the so-called Skip-gram model in Word2vec. However, as a bag-of-words model, the skip-gram model mainly utilized the local structure information. The lack of information metrics for vertices in global network leads to the mix of vertices with different labels in the new embedding space. To solve this problem, in this paper we propose a Network Representation Learning method with Deep Metric Learning, namely DML-NRL. By setting the initialized anchor vertices and adding the similarity measure in the training progress, the distance information between different labels of vertices in the network is integrated into the vertex representation, which improves the accuracy of network embedding algorithm effectively. We compare our method with baselines by applying them to the tasks of multi-label classification and data visualization of vertices. The experimental results show that our method outperforms the baselines in all three datasets, and the method has proved to be effective and robust.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDP7233/_p

Copy

@ARTICLE{e102-d_3_568,
author={Xiaotao CHENG, Lixin JI, Ruiyang HUANG, Ruifei CUI, },
journal={IEICE TRANSACTIONS on Information},
title={Network Embedding with Deep Metric Learning},
year={2019},
volume={E102-D},
number={3},
pages={568-578},
abstract={Network embedding has attracted an increasing amount of attention in recent years due to its wide-ranging applications in graph mining tasks such as vertex classification, community detection, and network visualization. Network embedding is an important method to learn low-dimensional representations of vertices in networks, aiming to capture and preserve the network structure. Almost all the existing network embedding methods adopt the so-called Skip-gram model in Word2vec. However, as a bag-of-words model, the skip-gram model mainly utilized the local structure information. The lack of information metrics for vertices in global network leads to the mix of vertices with different labels in the new embedding space. To solve this problem, in this paper we propose a Network Representation Learning method with Deep Metric Learning, namely DML-NRL. By setting the initialized anchor vertices and adding the similarity measure in the training progress, the distance information between different labels of vertices in the network is integrated into the vertex representation, which improves the accuracy of network embedding algorithm effectively. We compare our method with baselines by applying them to the tasks of multi-label classification and data visualization of vertices. The experimental results show that our method outperforms the baselines in all three datasets, and the method has proved to be effective and robust.},
keywords={},
doi={10.1587/transinf.2018EDP7233},
ISSN={1745-1361},
month={March},}

Copy

TY - JOUR
TI - Network Embedding with Deep Metric Learning
T2 - IEICE TRANSACTIONS on Information
SP - 568
EP - 578
AU - Xiaotao CHENG
AU - Lixin JI
AU - Ruiyang HUANG
AU - Ruifei CUI
PY - 2019
DO - 10.1587/transinf.2018EDP7233
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 3
JA - IEICE TRANSACTIONS on Information
Y1 - March 2019
AB - Network embedding has attracted an increasing amount of attention in recent years due to its wide-ranging applications in graph mining tasks such as vertex classification, community detection, and network visualization. Network embedding is an important method to learn low-dimensional representations of vertices in networks, aiming to capture and preserve the network structure. Almost all the existing network embedding methods adopt the so-called Skip-gram model in Word2vec. However, as a bag-of-words model, the skip-gram model mainly utilized the local structure information. The lack of information metrics for vertices in global network leads to the mix of vertices with different labels in the new embedding space. To solve this problem, in this paper we propose a Network Representation Learning method with Deep Metric Learning, namely DML-NRL. By setting the initialized anchor vertices and adding the similarity measure in the training progress, the distance information between different labels of vertices in the network is integrated into the vertex representation, which improves the accuracy of network embedding algorithm effectively. We compare our method with baselines by applying them to the tasks of multi-label classification and data visualization of vertices. The experimental results show that our method outperforms the baselines in all three datasets, and the method has proved to be effective and robust.
ER -

IEICE TRANSACTIONS on Information