
For FullText PDF, please login, if you are a member of IEICE,
or go to Pay Per View on menu list, if you are a nonmember of IEICE.

Network Embedding with Deep Metric Learning
Xiaotao CHENG Lixin JI Ruiyang HUANG Ruifei CUI
Publication
IEICE TRANSACTIONS on Information and Systems
Vol.E102D
No.3
pp.568578 Publication Date: 2019/03/01
Online ISSN: 17451361
DOI: 10.1587/transinf.2018EDP7233
Type of Manuscript: PAPER Category: Artificial Intelligence, Data Mining Keyword: deep metric learning, network representation learning, likelihood label, anchor initialization, semisupervised learning,
Full Text: PDF(1.7MB) >>Buy this Article
Summary:
Network embedding has attracted an increasing amount of attention in recent years due to its wideranging applications in graph mining tasks such as vertex classification, community detection, and network visualization. Network embedding is an important method to learn lowdimensional representations of vertices in networks, aiming to capture and preserve the network structure. Almost all the existing network embedding methods adopt the socalled Skipgram model in Word2vec. However, as a bagofwords model, the skipgram 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 DMLNRL. 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 multilabel 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.

