计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (11): 62-70.doi: 10.11896/jsjkx.220900166

• 数据库&大数据&数据科学 • 上一篇    下一篇

基于对比学习的多关系属性图聚类方法

谢卓1, 康乐2, 周丽娟1, 张志鸿1   

  1. 1 郑州大学计算机与人工智能学院 郑州 450000
    2 清华大学计算机科学与技术系 北京 100000
  • 收稿日期:2022-09-17 修回日期:2022-12-05 出版日期:2023-11-15 发布日期:2023-11-06
  • 通讯作者: 周丽娟(ieljzhou@zzu.edu.cn)
  • 作者简介:(zhuoxiexz@163.com)
  • 基金资助:
    国家自然科学基金(62006211);河南省重大公益专项(201300210300)

Clustering Method Based on Contrastive Learning for Multi-relation Attribute Graph

XIE Zhuo1, KANG Le2, ZHOU Lijuan1, ZHANG Zhihong1   

  1. 1 School of Computer and Artificial Intelligence,Zhengzhou University,Zhengzhou 450000,China
    2 Department of Computer Science and Technology,Tsinghua University,Beijing 100000,China
  • Received:2022-09-17 Revised:2022-12-05 Online:2023-11-15 Published:2023-11-06
  • About author:XIE Zhuo,born in 1997,postgraduate,is a member of China Computer Federation.His main research interestis deep learning on graphs.ZHOU Lijuan,born in 1987,Ph.D,associate professor,is a member of China Computer Federation.Her main research interests include cross modal semantic understanding and computer vision.
  • Supported by:
    National Natural Science Foundation of China(62006211) and Major Public Welfare Projects in Henan Province(201300210300).

PDF (PC)

4496

摘要： 现实世界包含复杂的图数据,其节点之间通常包含多种关系,这种图被称为多关系属性图。图聚类是挖掘图数据相似信息的技术之一,然而现有的图聚类的方法大多只适用于单关系图。即使有的方法考虑到了多关系图,也往往是将图表示学习与聚类看作两个单独的过程。受Deep Graph Infomax(DGI)算法的启发,文中设计了一种基于对比学习的多关系属性图的聚类方法(CCLMAG),用于解决上述问题:1)通过引入社区级互信息机制,弥补了DGI算法无法融合簇信息的缺点;2)引入嵌入融合模块来聚合不同关系上的节点嵌入;3)引入聚类优化模块将图表示学习与聚类两个过程联系起来,使得学习到的节点表示更适合聚类任务。在3个公开数据集和1个构建的期货数据集上的大量实验表明,所提方法优于目前最先进的基线方法,且具有实际应用价值。

关键词: 聚类, 多关系属性图, 图对比学习, 图表示学习, 无监督学习

Abstract: In the real world,there are many complex graph data which includes multiple relations between nodes,namely multi-relation attribute graph.Graph clustering is one of the approaches for mining similar information from graph data.However,most existing graph clustering methods assume that only single type of relation exists between nodes.Even for those that considering the multi-relation of a graph,they use only node attributes for training,or regard graph representation learning and clustering as two completely independent processes.Recently,Deep Graph Infomax(DGI) has shown promising results on many downstream tasks.But there are two major limitations for DGI.Firstly,DGI does not fully explore the various relations among nodes.Secondly,DGI does not jointly optimize the graph representation learning and clustering tasks,resulting in suboptimal clustering results.To address the above-mentioned problems,this paper proposes a novel framework,called clustering method based on contrastive learning for multi-relation attribute graph(CCLMAG),for learning the node embedding suitable for clustering in a unsupervised way.To be more specific,1)The community-level mutual information mechanism is applied to solve the problem of ignoring cluster information by DGI;2)the Embedding Fusion Module is augmented to aggregate the embedding of nodes in different relationships;3)the clustering optimization module is added to link the graph representation learning and clustering so that the learned node representation is more suitable for the clustering task,thus enhancing the interpretability of the clustering results.Extensive experimental results on three multi-relation attribute graph datasets and a real-world futures dataset demonstrate the superiority of CCLMAG compared with the state-of-the-art methods.

Key words: Clustering, Multi-relation attribute graph, Graph contrastive learning, Graph representation learning, Unsupervised learning

中图分类号: 

  • TP391
[1]SU X,XUE S,LIU F Z,et al.A Comprehensive Survey on Community Detection With Deep Learning[J].arXiv:2105.12584,2021.
[2]HARTIGAN J A,WONG M.Algorithm AS 136:A k-means clustering algorithm[J].Journal of the Royal Statistical Society(Series C),1979,28(1):100-108.
[3]ANDREW N,MICHAEL J,YAIR W.On spectral clustering:Analysis and an algorithm[C]//Advances in Neural Information Processing Systems 14.2001:849-856.
[4]MARTIN E,HANS-PETER K,SANDER J,et al.A density-based algorithm for discovering clusters in large spatial databa-ses with noise[C]//Proceedings of the Second International Conference on Knowledge Discovery and Data Mining.1996:226-231.
[5]WANG T,YANG G Y,WU J H,et al.Dual Contrastive Attri-buted Graph Clustering Network[J].arXiv:2206.07897,2022.
[6]PETAR V,WILLIAM F,WILLIAM L,et al.Deep Graph Infomax[J].arXiv:1809.10341,2018.
[7]BU Z,WANG Y Y,MA L N,et al.Attributed Graph Clustering Approach Based on Dynamic Cluster Formation Game[J].Chinese Journal of Computers,2021,44(9):1824-1840.
[8]FAN S H,WANG X,SHI C,et al.One2Multi Graph Autoencoder for Multi-view Graph Clustering[C]//Proceedings of The Web Conference 2020.2020:3070-3076.
[9]SHI Y,HAN F Q,HE X W,et al.mvn2vec:Preservation and collaboration in multi-view network embedding[J].arXiv:1801.06597,2018.
[10]WANG X,JI H Y,SHI C,et al.Heterogeneous Graph Attention Network[C]//The World Wide Web Conference.2019:2022-2032.
[11]MICHAEL S,THOMAS N K,PETER B,et al.Modeling relational data with graph convolutional networks[C]//European Semantic Web Conference 2018.2018:593-607.
[12]PARK C Y,KIM D,HAN J W,et al.Unsupervised Attributed Multiplex Network Embedding[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2020:5371-5378.
[13]WANG J,CAO J,LI W,et al.CANE:community-aware network embedding via adversarial training[C]//Knowledge and Information Systems.2021:411-438.
[14]KANG Y,KOU Y Q,XIE S Y,et al.Deep Clustering ModelBased on Fusion Variational Graph Attention Self-encoder [J].Computer Science,2021,48(S2):81-87,116.
[15]YANG X H,WANG L,YE L,et al.Complex Network Commu-nity Detection Algorithm Based on Node Similarity and Network Embedding [J].Computer Science,2022,49(3):121-128.
[16]PAN S Y,QIN Y M,LI T Y,et al.Momentum contrast for unsupervised visual representation learning[C]//CVF Conference on Computer Vision and Pattern Recognition Workshops.2022:2646-2655.
[17]LU S H,CHEN L,LU G Y,et al.Weakly-Supervised Contrastive Learning Framework for Few-Shot Sentiment Classification Tasks [J].Journal of Computer Research and Development,2022,59(9):2003-2014.
[18]WANG X,LIU N,HAN H,et al.Self-supervised Heteroge-neous Graph Neural Network with Co-contrastive Learning[C]//Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining.2021:1726-1736.
[19]ZHAO H,YANG X,WANG Z R,et al.Graph Debiased Con-trastive Learning with Joint Representation Clustering[C]//International Joint Conferences on Artificial Intelligence Organization.2021:3434-3440.
[20]YOU Q Z,JIN H L,WANG Z W,et al.Image captioning with semantic attention[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition.2016:4651-4659.
[21]YANG Y H,WU L,HONG R C,et al.Enhanced Graph Lear-ning for Collaborative Filtering via Mutual Information Maximization[C]//Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval.2021:71-80.
[22]BRYAN P,RAMI A,STEVEN S,et al.Deepwalk:Online lear-ning of social representations[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Disco-very and Data Mining.2014:701-710.
[23]ADITYA G,JURE L.node2vec:Scalable feature learning fornetworks[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2016:855-864.
[24]TANG J,QU M,WANG M Z,et al.Line:Large-scale information network embedding[C]//Proceedings of the 24th International Conference on World Wide Web.2015:1067-1077.
[25]THOMAS K M W.Semi-Supervised Classification with GraphConvolutional Networks[J].arXiv:1609.02907,2016.
[26]JING B Y,PARK C Y,TONG H H,et al.HDMI:High-orderDeep Multiplex Infomax[C]//Proceedings of the Web Confe-rence 2021.2021:2414-2424.
[27]BO D Y,WANG X,SHI C,et al.Structural deep clustering network[C]//Proceedings of The Web Conference 2020.2020:1400-1410.
[28]DIEDERIK P K,JIMMY B.Adam:A method for stochastic optimization[J].arXiv:1412.6980,2014.
[29]LAURENS V D M,GEOFFREY H.Visualizing data usingt-SNE[C]//Journal of Machine Learning Research.2008:2579-2605.
[30]WANG H B,ZHOU C,WU J,et al.Deep Structure Learning for Fraud Detection[C]//2018 IEEE International Conference on Data Mining.2018:567-576.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发