Article(id=1325493590348145109, tenantId=1146029695717560320, journalId=1324759914146529478, issueId=1325493569112388051, articleNumber=null, orderNo=null, doi=10.7544/issn1000-1239.202550356, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=research-article, receivedDate=1748188800000, receivedDateStr=2025-05-26, revisedDate=1755446400000, revisedDateStr=2025-08-18, acceptedDate=1756742400000, acceptedDateStr=2025-09-02, onlineDate=1793951898419, onlineDateStr=2026-11-06, pubDate=1769875200000, pubDateStr=2026-02-01, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1793951898419, onlineIssueDateStr=2026-11-06, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1793951898419, creator=13701087609, updateTime=1793951898419, updator=13701087609, issue=Issue{id=1325493569112388051, tenantId=1146029695717560320, journalId=1324759914146529478, year='2026', volume='63', issue='2', pageStart='275', pageEnd='538', issueExtLink='null', onlineDate='null', pubDate='1769875200000', pubDateStr='2026-02-01', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1793951893355, creator='13701087609', updateTime=1793952511329, updator='13701087609', preIssue=null, nextIssue=null, articleTotal=null, ext={EN=IssueExt(id=1325496161167090054, tenantId=1146029695717560320, journalId=1324759914146529478, issueId=1325493569112388051, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1325496161167090055, tenantId=1146029695717560320, journalId=1324759914146529478, issueId=1325493569112388051, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null, downloadFileDto=null}, startPage=324, endPage=337, ext={EN=ArticleExt(id=1325493593376432599, articleId=1325493590348145109, tenantId=1146029695717560320, journalId=1324759914146529478, language=EN, title=Motif-Aware Multi-View Cooperative Clustering Optimization Algorithm, columnId=1325489891789910943, journalTitle=Journal of Computer Research and Development, columnName=Artificial Intelligence, runingTitle=null, highlight=null, articleAbstract=

Graph neural networks (GNNs) learn graph embeddings by iteratively aggregating neighborhood features and have been widely applied in graph data analysis. Existing methods primarily focus on low-order node-edge interactions while overlooking high-order group interaction patterns captured by motifs. This limitation hinders the effective modeling of high-order dependencies in complex networks. Motifs, as frequently occurring functional substructures in networks, can effectively reveal high-order semantic relationships among nodes. The motif co-occurrence view provides a novel perspective for representing such interactions. However, the weak connectivity of motif co-occurrence views restricts message passing in GNNs, thereby impeding the effective propagation of global information. To address this issue, we propose a motif-aware multi-view cooperative clustering optimization algorithm (MMCC), which fully exploits high-order topological information through an adaptive multi-view fusion mechanism while enhancing representation consistency across different views via contrastive learning to alleviate the issue of restricted message passing. Specifically, MMCC first constructs multiple motif co-occurrence views based on different motif structures and employs a semantic attention-based multi-view auto-encoder to dynamically learn the importance of different motif views, achieving adaptive fusion. Then, contrastive learning is introduced to enforce embedding space consistency between the original view and motif co-occurrence views, mitigating the message passing limitations caused by weak connectivity. Finally, by optimizing a KL-divergence-based objective function, MMCC jointly optimizes feature learning and clustering tasks. Experimental results on seven real-world networks demonstrate that MMCC outperforms eight baseline algorithms in clustering accuracy (ACC), normalized mutual information (NMI), F1-score (F1), and adjusted Rand index (ARI), highlighting its effectiveness in high-order network clustering.

, authors=Meilin Liu1, Hua Li1, Wenping Zheng1, 2, 3, *, authorsList=Meilin Liu, Hua Li, Wenping Zheng, authorCompany=null, correspAuthors=Wenping Zheng, authorNote=

born in 1996. PhD candidate. Student member of CCF. Her main research interest includes complex network analysis

born in 2001. Master candidate. Student member of CCF. Her main research interest includes complex network analysis

born in 1979. PhD, professor, PhD supervisor. Member of CCF. Her main research interests include complex network analysis and bioinformatics

, correspAuthorsNote=null, copyrightStatement=Copyright © 2026 Journal of Computer Research and Development. All rights reserved., copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, fund=null), CN=ArticleExt(id=1325493618449981943, articleId=1325493590348145109, tenantId=1146029695717560320, journalId=1324759914146529478, language=CN, title=模体感知的多视图协同聚类优化算法, columnId=1325489892104483745, journalTitle=计算机研究与发展, columnName=人工智能, runingTitle=null, highlight=null, articleAbstract=

图神经网络通过迭代聚合邻域特征学习图的嵌入表示,已广泛应用于图数据分析。现有方法主要关注低阶点边交互,而对以模体为载体的高阶成组交互模式关注不足,导致复杂网络中的高阶依赖关系难以被充分捕捉。模体作为网络中频繁出现的功能性子结构,能够有效揭示节点间的高阶语义关联,而模体共现视图则为刻画此类交互模式提供了新的表征视角。然而,模体共现视图的弱连通性限制了图神经网络的消息传递能力,影响全局信息的有效传播。针对此提出模体感知的多视图协同聚类优化算法(motif-aware multi-view cooperative clustering optimization algorithm,MMCC),通过自适应多视图融合机制充分挖掘高阶拓扑信息,同时利用对比学习增强不同视图间的表征一致性,从而缓解消息传递受限问题。具体而言,首先MMCC基于不同模体构建多个模体共现视图,并设计基于语义注意力的多视图自编码器动态学习不同模体视图的重要性,实现各视图的自适应融合;其次,引入对比学习约束原始视图与模体共现视图的嵌入空间一致性,缓解因模体共现视图弱连通性导致的消息传递受限问题;最后,通过优化基于KL散度的目标函数,实现特征学习与聚类任务的联合优化。在7个真实网络数据集上的聚类结果表明,MMCC在准确率(accuracy,ACC)、标准化互信息(normalized mutual information,NMI)、F1分数(F1)和调整兰德系数(adjusted Rand index,ARI)上较8个基准算法展现更显著的优势。

, authors=刘美麟1, 李华1, 郑文萍1, 2, 3, *, authorsList=刘美麟, 李华, 郑文萍, authorCompany=null, correspAuthors=郑文萍, authorNote=

1996年生。博士研究生。CCF学生会员。主要研究方向为复杂网络分析

2001年生。硕士研究生。CCF学生会员。主要研究方向为复杂网络分析

1979年生。博士,教授,博士生导师。CCF会员。主要研究方向为复杂网络分析、生物信息学

, correspAuthorsNote=
郑文萍()
, copyrightStatement=版权所有 © 2026 计算机研究与发展编辑部, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=Rff6F0r62XMyh7gyXh2hLQ==, magXml=A78oe0bHVPRgtcaWrCR1NA==, pdfUrl=null, pdf=jwPZQkAwmWhmANCi6L1mMw==, pdfFileSize=2607554, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=DPG7fZd4Wu5CH50xg8Ikqw==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=3P/J7nC9UFstqFXyNNMJ3g==, mapNumber=null, fund=null)}, authors=[Author(id=1325493623881605634, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=834754720@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1325493624401699332, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493623881605634, language=EN, stringName=Meilin Liu, firstName=Meilin, middleName=null, lastName=Liu, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006, bio={"content":"

born in 1996. PhD candidate. Student member of CCF. Her main research interest includes complex network analysis

"}, bioImg=null, bioContent=

born in 1996. PhD candidate. Student member of CCF. Her main research interest includes complex network analysis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1325493626448519685, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493623881605634, language=CN, stringName=刘美麟, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1山西大学计算机与信息技术学院 太原 030006, bio={"img":"l7LqkR1upwg0T4salU033A==","content":"

1996年生。博士研究生。CCF学生会员。主要研究方向为复杂网络分析

"}, bioImg=l7LqkR1upwg0T4salU033A==, bioContent=

1996年生。博士研究生。CCF学生会员。主要研究方向为复杂网络分析

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1325493620219978232, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=1, ext=[AuthorCompanyExt(id=1325493620526162425, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493620555522554, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1山西大学计算机与信息技术学院 太原 030006)])]), Author(id=1325493626821812743, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1325493628977685001, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493626821812743, language=EN, stringName=Hua Li, firstName=Hua, middleName=null, lastName=Li, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006, bio={"content":"

born in 2001. Master candidate. Student member of CCF. Her main research interest includes complex network analysis

"}, bioImg=null, bioContent=

born in 2001. Master candidate. Student member of CCF. Her main research interest includes complex network analysis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1325493629350978058, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493626821812743, language=CN, stringName=李华, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1山西大学计算机与信息技术学院 太原 030006, bio={"img":"tKZXn05ELG0b6JMTEuEAJQ==","content":"

2001年生。硕士研究生。CCF学生会员。主要研究方向为复杂网络分析

"}, bioImg=tKZXn05ELG0b6JMTEuEAJQ==, bioContent=

2001年生。硕士研究生。CCF学生会员。主要研究方向为复杂网络分析

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1325493620219978232, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=1, ext=[AuthorCompanyExt(id=1325493620526162425, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493620555522554, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1山西大学计算机与信息技术学院 太原 030006)])]), Author(id=1325493631024505356, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=wpzheng@sxu.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1325493632689644048, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493631024505356, language=EN, stringName=Wenping Zheng, firstName=Wenping, middleName=null, lastName=Zheng, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, *, address=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006
2Key Laboratory of Computation Intelligence and Chinese Information Processing (Shanxi University), Ministry of Education, Taiyuan 030006
3Institute of Intelligent Information Processing, Shanxi University, Taiyuan 030006, bio={"content":"

born in 1979. PhD, professor, PhD supervisor. Member of CCF. Her main research interests include complex network analysis and bioinformatics

"}, bioImg=null, bioContent=

born in 1979. PhD, professor, PhD supervisor. Member of CCF. Her main research interests include complex network analysis and bioinformatics

, aboutCorrespAuthor=null), CN=AuthorExt(id=1325493633134240273, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493631024505356, language=CN, stringName=郑文萍, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, *, address=1山西大学计算机与信息技术学院 太原 030006
2计算智能与中文信息处理教育部重点实验室(山西大学) 太原 030006
3山西大学智能信息处理研究所 太原 030006, bio={"img":"x2PFQ7SpDLkSYh1vM+vbIg==","content":"

1979年生。博士,教授,博士生导师。CCF会员。主要研究方向为复杂网络分析、生物信息学

"}, bioImg=x2PFQ7SpDLkSYh1vM+vbIg==, bioContent=

1979年生。博士,教授,博士生导师。CCF会员。主要研究方向为复杂网络分析、生物信息学

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1325493620219978232, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=1, ext=[AuthorCompanyExt(id=1325493620526162425, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493620555522554, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1山西大学计算机与信息技术学院 太原 030006)]), AuthorCompany(id=1325493623067910651, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=2, ext=[AuthorCompanyExt(id=1325493623076299260, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623067910651, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Key Laboratory of Computation Intelligence and Chinese Information Processing (Shanxi University), Ministry of Education, Taiyuan 030006), AuthorCompanyExt(id=1325493623084687869, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623067910651, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2计算智能与中文信息处理教育部重点实验室(山西大学) 太原 030006)]), AuthorCompany(id=1325493623168573950, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=3, ext=[AuthorCompanyExt(id=1325493623176962559, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623168573950, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3Institute of Intelligent Information Processing, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493623185351168, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623168573950, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3山西大学智能信息处理研究所 太原 030006)])])], keywords=[Keyword(id=1325493634010849811, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, orderNo=1, keyword=complex network), Keyword(id=1325493635659211284, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, orderNo=2, keyword=motif), Keyword(id=1325493636883948054, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, orderNo=3, keyword=high-order interaction), Keyword(id=1325493637752168983, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, orderNo=4, keyword=graph neural network), Keyword(id=1325493639840932376, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, orderNo=5, keyword=contrastive learning), Keyword(id=1325493639941595673, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, orderNo=6, keyword=attention mechanism), Keyword(id=1325493640675598874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, orderNo=1, keyword=复杂网络), Keyword(id=1325493641170526747, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, orderNo=2, keyword=模体), Keyword(id=1325493641568985628, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, orderNo=3, keyword=高阶相互作用), Keyword(id=1325493641963250205, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, orderNo=4, keyword=图神经网络), Keyword(id=1325493642797916702, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, orderNo=5, keyword=对比学习), Keyword(id=1325493644450472479, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, orderNo=6, keyword=注意力机制)], refs=[Reference(id=1325493665010950718, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=1, rfOrder=0, authorNames=null, journalName=null, refType=null, unstructuredReference=Watts D J, Strogatz S H. Collective dynamics of ‘small-world’ networks[J]. Nature, 1998, 393(6684): 440−442, articleTitle=null, refAbstract=null), Reference(id=1325493666705449535, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=2, rfOrder=1, authorNames=null, journalName=null, refType=null, unstructuredReference=Girvan M, Newman M E J. Community structure in social and biological networks[J]. Proceedings of the National Academy of Science, 2002, 99(12): 7821−7826, articleTitle=null, refAbstract=null), Reference(id=1325493667133268544, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=3, rfOrder=2, authorNames=null, journalName=null, refType=null, unstructuredReference=Perozzi B, Al-Rfou R, Skiena S. DeepWalk: Online learning of social representations[C]//Proc of the 20th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2014: 701−710, articleTitle=null, refAbstract=null), Reference(id=1325493667519144513, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=4, rfOrder=3, authorNames=null, journalName=null, refType=null, unstructuredReference=Grover A, Leskovec J. node2vec: Scalable feature learning for Networks[C]//Proc of the 22nd ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2016: 855−864, articleTitle=null, refAbstract=null), Reference(id=1325493667930186306, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=5, rfOrder=4, authorNames=null, journalName=null, refType=null, unstructuredReference=Kipf T N, Welling M. Semi-supervised classification with graph convolutional networks[C/OL]//Proc of the 5th Int Conf on Learning Representations. Washington: ICLR, 2017[2025-06-26]. https://openreview.net/forum?id=SJU4ayYgl, articleTitle=null, refAbstract=null), Reference(id=1325493668345422403, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=6, rfOrder=5, authorNames=null, journalName=null, refType=null, unstructuredReference=Milo R, Shen-Orr S, Itzkovitz S, et al. Network motifs: Simple building blocks of complex networks[J]. Science, 2002, 298(5594): 824−827, articleTitle=null, refAbstract=null), Reference(id=1325493668769047108, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=7, rfOrder=6, authorNames=null, journalName=null, refType=null, unstructuredReference=Xia Feng, Yu Shuo, Liu Chengfei, et al. CHIEF: Clustering with higher-order motifs in big networks[J]. IEEE Transactions on Network Science and Engineering, 2022, 9(3): 990−1005, articleTitle=null, refAbstract=null), Reference(id=1325493669184283205, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=8, rfOrder=7, authorNames=null, journalName=null, refType=null, unstructuredReference=Benson A R, Gleich D F, Leskovec J. Higher-order organization of complex networks[J]. Science, 2016, 353(6295): 163−166, articleTitle=null, refAbstract=null), Reference(id=1325493669662433862, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=9, rfOrder=8, authorNames=null, journalName=null, refType=null, unstructuredReference=Wernicke S. Efficient detection of network motifs[J]. IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2006, 3(4): 347−359, articleTitle=null, refAbstract=null), Reference(id=1325493671445013063, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=10, rfOrder=9, authorNames=null, journalName=null, refType=null, unstructuredReference=Ribeiro P, Silva F. G-Tries: An efficient data structure for discovering network motifs[C]//Proc of the 2010 ACM Symp on Applied Computing. New York: ACM, 2010: 1559−1566, articleTitle=null, refAbstract=null), Reference(id=1325493671755391560, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=11, rfOrder=10, authorNames=null, journalName=null, refType=null, unstructuredReference=Zhao Peixiang. gSparsify: Graph motif based sparsification for graph clustering[C]//Proc of the 24th ACM Int Conf on Information and Knowledge Management. New York: ACM, 2015: 373−382, articleTitle=null, refAbstract=null), Reference(id=1325493672124490313, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=12, rfOrder=11, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Peizhen, Huang Ling, Wang Changdong, et al. EdMot: An edge enhancement approach for motif-aware community detection[C]//Proc of the 25th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2019: 479−487, articleTitle=null, refAbstract=null), Reference(id=1325493672577475146, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=13, rfOrder=12, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Pan, Dau H, Puleo G, et al. Motif clustering and overlapping clustering for social network analysis[C/OL]//Proc of the 2017 IEEE Conf on Computer Communications. Piscataway, NJ: IEEE, 2017[2025-06-26]. https://ieeexplore.ieee.org/document/8056956, articleTitle=null, refAbstract=null), Reference(id=1325493673005294155, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=14, rfOrder=13, authorNames=null, journalName=null, refType=null, unstructuredReference=Hou Zhenyu, Liu Xiao, Cen Yukuo, et al. GraphMAE: Self-supervised masked graph autoencoders[C]//Proc of the 28th ACM SIGKDD Conf on Knowledge Discovery and Data Mining. New York: ACM, 2022: 594−604, articleTitle=null, refAbstract=null), Reference(id=1325493673403753037, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=15, rfOrder=14, authorNames=null, journalName=null, refType=null, unstructuredReference=马帅, 刘建伟, 左信. 图神经网络综述[J]. 计算机研究与发展, 2022, 59(1): 47−80, articleTitle=null, refAbstract=null), Reference(id=1325493673852543566, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=15, rfOrder=15, authorNames=null, journalName=null, refType=null, unstructuredReference=Ma Shuai, Liu Jianwei, Zuo Xin. Survey on graph neural network[J]. Journal of Computer Research and Development, 2022, 59(1): 47−80 (in Chinese), articleTitle=null, refAbstract=null), Reference(id=1325493675526070863, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=16, rfOrder=16, authorNames=null, journalName=null, refType=null, unstructuredReference=Velickovic P, Cucurull G, Casanova A, et al. Graph attention networks[C/OL]//Proc of the 6th Int Conf on Learning Representations. Washington: ICLR. 2018[2025-06-26]. https://openreview.net/forum?id=rJXMpikCZ, articleTitle=null, refAbstract=null), Reference(id=1325493676771779152, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=17, rfOrder=17, authorNames=null, journalName=null, refType=null, unstructuredReference=You Yunning, Chen Tianlong, Sui Yongduo, et al. Graph contrastive learning with augmentations[C]//Proc of the 34th Advances in Neural Information Processing Systems. Cambridge, MA: MIT, 2020, 5812−5823, articleTitle=null, refAbstract=null), Reference(id=1325493677606445649, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=18, rfOrder=18, authorNames=null, journalName=null, refType=null, unstructuredReference=Qiang Qianyao, Zhang Bin, Wang Fei, et al. Fast multi-view discrete clustering with anchor graphs[C]//Proc of the 35th AAAI Conf on Artificial Intelligence. Palo Alto, CA: AAAI, 2021: 9360−9367, articleTitle=null, refAbstract=null), Reference(id=1325493679330304594, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=19, rfOrder=19, authorNames=null, journalName=null, refType=null, unstructuredReference=Kipf T N, Welling M. Variational graph auto-encoders[J]. arXiv preprint, arXiv: 1611.07308, 2016, articleTitle=null, refAbstract=null), Reference(id=1325493680571818581, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=20, rfOrder=20, authorNames=null, journalName=null, refType=null, unstructuredReference=Wang Chun, Pan Shirui, Hu Ruiqi, et al. Attributed graph clustering: A deep attentional embedding approach[C]//Proc of the 28th Int Joint Conf on Artificial Intelligence. San Francisco, CA: Morgan Kaufmann, 2019: 3670−3676, articleTitle=null, refAbstract=null), Reference(id=1325493683088400982, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=21, rfOrder=21, authorNames=null, journalName=null, refType=null, unstructuredReference=Bo Deyu, Wang Xiao, Shi Chuan, et al. Structural deep clustering network[C]//Proc of the 29th Int Conf on World Wide Web. New York: ACM, 2020: 1400−1410, articleTitle=null, refAbstract=null), Reference(id=1325493685198135895, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=22, rfOrder=22, authorNames=null, journalName=null, refType=null, unstructuredReference=Huo Guangyu, Zhang Yong, Gao Junbin, et al. CaEGCN: Cross-attention fusion based enhanced graph convolutional network for clustering[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(4): 3471−3483, articleTitle=null, refAbstract=null), Reference(id=1325493687303676506, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=23, rfOrder=23, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Jintang, Wu Ruofan, Sun Wangbin, et al. What’s behind the mask: Understanding masked graph modeling for graph autoencoders[C]//Proc of the 29th ACM SIGKDD Conf on Knowledge Discovery and Data Mining. New York: ACM, 2023: 1268−1279, articleTitle=null, refAbstract=null), Reference(id=1325493689753150046, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=24, rfOrder=24, authorNames=null, journalName=null, refType=null, unstructuredReference=Bai Lu, Xu Zhuo, Cui Lixin, et al. HC-GAE: The hierarchical cluster-based graph auto-encoder for graph representation learning[C]//Proc of the 38th Advances in Neural Information Proc Systems. Cambridge, MA: MIT, 2024: 127968−127986, articleTitle=null, refAbstract=null), Reference(id=1325493691032412767, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=25, rfOrder=25, authorNames=null, journalName=null, refType=null, unstructuredReference=Yin Hao, Benson A R, Leskovec J, et al. Local higher-order graph clustering[C]//Proc of the 23rd ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2017: 555−564, articleTitle=null, refAbstract=null), Reference(id=1325493693964231265, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=26, rfOrder=26, authorNames=null, journalName=null, refType=null, unstructuredReference=Huang Ling, Wang Changdong, Chao Hongyang. HM-Modularity: A harmonic motif modularity approach for multi-layer network community detection[J]. IEEE Transactions on Knowledge and Data Engineering, 2021, 33(6): 2520−2533, articleTitle=null, refAbstract=null), Reference(id=1325493694400438882, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=27, rfOrder=27, authorNames=null, journalName=null, refType=null, unstructuredReference=Yu Yanlei, Lu Zhiwu, Liu Jiajun, et al. RUM: Network representation learning using motifs[C]//Proc of the 35th Int Conf on Data Engineering. Piscataway, NJ: IEEE, 2019: 1382−1393, articleTitle=null, refAbstract=null), Reference(id=1325493695193162339, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=28, rfOrder=28, authorNames=null, journalName=null, refType=null, unstructuredReference=Dareddy M R, Das M, Yang Hao. motif2vec: Motif aware node representation learning for heterogeneous networks[C]//Proc of the 2019 IEEE Int Conf on Big Data. Piscataway, NJ: IEEE, 2019: 1052−1059, articleTitle=null, refAbstract=null), Reference(id=1325493695679701605, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=29, rfOrder=29, authorNames=null, journalName=null, refType=null, unstructuredReference=Wang Lei, Ren Jing, Xu Bo, et al. MODEL: Motif-based deep feature learning for link prediction[J]. IEEE Transactions on Computational Social Systems, 2020, 7(2): 503−516, articleTitle=null, refAbstract=null), Reference(id=1325493696103326310, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=30, rfOrder=30, authorNames=null, journalName=null, refType=null, unstructuredReference=Shao Ping, Yang Yang, Xu Shengyao, et al. Network embedding via motifs[J]. ACM Transactions on Knowledge Discovery from Data, 2022, 16(3): 44: 1−44: 20, articleTitle=null, refAbstract=null), Reference(id=1325493696518562407, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=31, rfOrder=31, authorNames=null, journalName=null, refType=null, unstructuredReference=Rossi R A, Ahmed N K, Koh E, et al. A structural graph representation learning framework[C]//Proc of the 13th Int Conf on Web Search and Data Mining. New York: ACM, 2020: 483−491, articleTitle=null, refAbstract=null), Reference(id=1325493698141758056, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=32, rfOrder=32, authorNames=null, journalName=null, refType=null, unstructuredReference=Lee J B, Rossi R A, Kong X, et al. Graph convolutional networks with motif-based attention[C]//Proc of the 28th ACM Int Conf on Information and Knowledge Management. New York: ACM, 2019: 499−508, articleTitle=null, refAbstract=null), Reference(id=1325493698959647337, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=33, rfOrder=33, authorNames=null, journalName=null, refType=null, unstructuredReference=Sun Mingchen, Yang Mengduo, Li Yingji, et al. Structural-aware motif-based prompt tuning for graph clustering[J]. Information Sciences, 2023, 649: 119643, articleTitle=null, refAbstract=null), Reference(id=1325493699471352426, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=34, rfOrder=34, authorNames=null, journalName=null, refType=null, unstructuredReference=Wang Xiao, Zhu Meiqi, Bo Deyu, et al. AM-GCN: Adaptive multi-channel graph convolutional networks[C]//Proc of the 26th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2020: 1243−1253, articleTitle=null, refAbstract=null), Reference(id=1325493699899171435, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=35, rfOrder=35, authorNames=null, journalName=null, refType=null, unstructuredReference=Mccallum A K, Nigam K, Rennie J. et al. Automating the construction of Internet portals with machine learning[J]. Information Retrieval, 2000, 3: 127−163, articleTitle=null, refAbstract=null), Reference(id=1325493700301824620, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=36, rfOrder=36, authorNames=null, journalName=null, refType=null, unstructuredReference=Giles C L, Bollacker K D, Lawrence S. CiteSeer: An automatic citation indexing system[C]//Proc of the 3rd ACM conf on Digital libraries. New York: ACM, 1998: 89−98, articleTitle=null, refAbstract=null), Reference(id=1325493701119713901, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=37, rfOrder=37, authorNames=null, journalName=null, refType=null, unstructuredReference=Pei Hongbin, Wei Bingzhe, Chang K, et al. Geom-GCN: Geometric graph convolutional networks[C/OL]//Proc of the 8th Int Conf on Learning Representations. Washington: ICLR. 2020 [2025-06-26]. https://openreview.net/forum?id=S1e2agrFvS, articleTitle=null, refAbstract=null), Reference(id=1325493702830989935, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=38, rfOrder=38, authorNames=null, journalName=null, refType=null, unstructuredReference=李嘉欣. 基于局部模体的社区发现算法研究[D]. 太原: 山西大学, 2023, articleTitle=null, refAbstract=null), Reference(id=1325493703602741873, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=38, rfOrder=39, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Jiaxin. Research on community detection algorithm based on local motifs[D]. Taiyuan: Shanxi University, 2023 (in Chinese), articleTitle=null, refAbstract=null), Reference(id=1325493704626152050, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=39, rfOrder=40, authorNames=null, journalName=null, refType=null, unstructuredReference=Bressan M, Leucci S, Panconesi A. Motivo: Fast motif counting via succinct color coding and adaptive sampling[J]. Proceedings of the VLDB Endowment, 2019, 12(11): 1651−1663, articleTitle=null, refAbstract=null), Reference(id=1325493705720865395, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=40, rfOrder=41, authorNames=null, journalName=null, refType=null, unstructuredReference=Yang Chen, Luo Lailong, Lu Yuliang, et al. SGES: A general and space-efficient framework for graphlet counting in graph streams[C]//Proc of the 33rd ACM Int Conf Information and Knowledge Management. New York: ACM, 2024: 2817−2826, articleTitle=null, refAbstract=null)], funds=null, companyList=[AuthorCompany(id=1325493620219978232, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=1, ext=[AuthorCompanyExt(id=1325493620526162425, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493620555522554, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1山西大学计算机与信息技术学院 太原 030006)]), AuthorCompany(id=1325493623067910651, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=2, ext=[AuthorCompanyExt(id=1325493623076299260, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623067910651, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Key Laboratory of Computation Intelligence and Chinese Information Processing (Shanxi University), Ministry of Education, Taiyuan 030006), AuthorCompanyExt(id=1325493623084687869, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623067910651, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2计算智能与中文信息处理教育部重点实验室(山西大学) 太原 030006)]), AuthorCompany(id=1325493623168573950, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=3, ext=[AuthorCompanyExt(id=1325493623176962559, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623168573950, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3Institute of Intelligent Information Processing, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493623185351168, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623168573950, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3山西大学智能信息处理研究所 太原 030006)])], figs=[ArticleFig(id=1325493645826204193, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 1, caption=Instances of ${\cal M}_2 $, ${\cal M}_3 $ and ${\cal M}_4 $, figureFileSmall=v6Ik2YRKvMnXhSuDpgc2Zg==, figureFileBig=DPG7fZd4Wu5CH50xg8Ikqw==, tableContent=null), ArticleFig(id=1325493646212080162, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图1, caption= ${\cal M}_2 $,${\cal M}_3 $,${\cal M}_4 $实例, figureFileSmall=v6Ik2YRKvMnXhSuDpgc2Zg==, figureFileBig=DPG7fZd4Wu5CH50xg8Ikqw==, tableContent=null), ArticleFig(id=1325493649064206883, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 2, caption=Framework of the motif-aware multi-view collaborative clustering optimization algorithm, figureFileSmall=23COOi3r5/kWcjE/y9074g==, figureFileBig=IGb7ft9nPYIXJOdn8ypp5g==, tableContent=null), ArticleFig(id=1325493649164870180, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图2, caption=模体感知的多视图协同聚类优化算法框架, figureFileSmall=23COOi3r5/kWcjE/y9074g==, figureFileBig=IGb7ft9nPYIXJOdn8ypp5g==, tableContent=null), ArticleFig(id=1325493649567523365, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 3, caption=Example of motif common view construction, figureFileSmall=PFIx2znhaT+tLG9PniD/hw==, figureFileBig=HJMuS4tc1UT3yAHK6imIOg==, tableContent=null), ArticleFig(id=1325493650368635431, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图3, caption=模体共现视图构建实例, figureFileSmall=PFIx2znhaT+tLG9PniD/hw==, figureFileBig=HJMuS4tc1UT3yAHK6imIOg==, tableContent=null), ArticleFig(id=1325493651169747496, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 4, caption=Example of weak connectivity in high-order network, figureFileSmall=85VYMzgB11TinKBdootghQ==, figureFileBig=X1Z80vxIEblgjtJA9K05Iw==, tableContent=null), ArticleFig(id=1325493651694035497, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图4, caption=高阶网络弱连通性实例, figureFileSmall=85VYMzgB11TinKBdootghQ==, figureFileBig=X1Z80vxIEblgjtJA9K05Iw==, tableContent=null), ArticleFig(id=1325493653304648234, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 5, caption=Ablation study on ACM and Cornell datasets, figureFileSmall=E9XQ4NSdQN1PiYxs9NOa/Q==, figureFileBig=QTrLHKARRNISW/CWrPefkw==, tableContent=null), ArticleFig(id=1325493654080594475, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图5, caption=ACM和Cornell数据集上的消融实验, figureFileSmall=E9XQ4NSdQN1PiYxs9NOa/Q==, figureFileBig=QTrLHKARRNISW/CWrPefkw==, tableContent=null), ArticleFig(id=1325493654554550828, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 6, caption=Distribution of attention coefficients among different motifs, figureFileSmall=fsvCbwKY1UjUAAVaTThGTA==, figureFileBig=g+C2HwXN+TjDRNWzrM4pBQ==, tableContent=null), ArticleFig(id=1325493654919455277, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图6, caption=不同模体的注意力系数分布, figureFileSmall=fsvCbwKY1UjUAAVaTThGTA==, figureFileBig=g+C2HwXN+TjDRNWzrM4pBQ==, tableContent=null), ArticleFig(id=1325493655036895790, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 7, caption=Motif abundance distribution based on BA network, figureFileSmall=Y3EXR1p8V8Nro5u3RA2W4w==, figureFileBig=5mW4Y66GAldIGdyvSUmzEw==, tableContent=null), ArticleFig(id=1325493655418577455, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图7, caption=基于BA网络的模体丰度分布, figureFileSmall=Y3EXR1p8V8Nro5u3RA2W4w==, figureFileBig=5mW4Y66GAldIGdyvSUmzEw==, tableContent=null), ArticleFig(id=1325493655787676208, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 8, caption=Pairwise similarity between motif features, figureFileSmall=yaHcfS5zaMW0XIINRgb6iQ==, figureFileBig=gGZPefZlDnkLbwDWqyT3lA==, tableContent=null), ArticleFig(id=1325493655863173681, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图8, caption=模体特征间的成对相似性, figureFileSmall=yaHcfS5zaMW0XIINRgb6iQ==, figureFileBig=gGZPefZlDnkLbwDWqyT3lA==, tableContent=null), ArticleFig(id=1325493656194523698, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 9, caption=Relative training time amplification of MMCC-4 compared with MMCC-3, figureFileSmall=Qo1JYSVTpjy5UChRdCKXwA==, figureFileBig=f9dKYmBt1DS1l5bSXhyw8g==, tableContent=null), ArticleFig(id=1325493657893216819, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图9, caption=MMCC-4相较MMCC-3的相对训练时间增幅, figureFileSmall=Qo1JYSVTpjy5UChRdCKXwA==, figureFileBig=f9dKYmBt1DS1l5bSXhyw8g==, tableContent=null), ArticleFig(id=1325493658375561780, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 10, caption=Impact of different embedding dimensions on performance, figureFileSmall=1c+IjNa5g6E9M2dSSRYRVg==, figureFileBig=yvFVBPBwiET1/DQYfmN/Ug==, tableContent=null), ArticleFig(id=1325493659143119413, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图10, caption=不同嵌入维度对性能的影响, figureFileSmall=1c+IjNa5g6E9M2dSSRYRVg==, figureFileBig=yvFVBPBwiET1/DQYfmN/Ug==, tableContent=null), ArticleFig(id=1325493659562549814, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Figure 11, caption=Impact of different learning rates on performance, figureFileSmall=ydqb8nA03PqZQQQeP+09EA==, figureFileBig=mdHpxfFsDICRSGYJZomVuA==, tableContent=null), ArticleFig(id=1325493660015534647, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=图11, caption=不同学习率对性能的影响, figureFileSmall=ydqb8nA03PqZQQQeP+09EA==, figureFileBig=mdHpxfFsDICRSGYJZomVuA==, tableContent=null), ArticleFig(id=1325493660485296696, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Table 1, caption=Statistics of the Datasets, figureFileSmall=null, figureFileBig=null, tableContent=
数据集节点数边数特征维度类别数平均度
ACM302513128187038.68
Cora27085429143374.00
DBLP4057352833441.74
CiteSeer33274552370362.74
Cornell195280170352.87
Texas183295170353.22
Wisconsin262466170353.56
), ArticleFig(id=1325493662485979705, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=表1, caption=

数据集统计

, figureFileSmall=null, figureFileBig=null, tableContent=
数据集节点数边数特征维度类别数平均度
ACM302513128187038.68
Cora27085429143374.00
DBLP4057352833441.74
CiteSeer33274552370362.74
Cornell195280170352.87
Texas183295170353.22
Wisconsin262466170353.56
), ArticleFig(id=1325493662901215802, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Table 2, caption=Index Value Comparative Results of MMCC and Benchmark Algorithms, figureFileSmall=null, figureFileBig=null, tableContent=
数据集指标EdMotmotif2vecGAESDCNCaEGCNDAEGCAGCNMPTGCMMCC-3
(本文)
MMCC-4
(本文)
注:黑体数值表示最优结果,下划线数值表示次优结果。
ACMACC0.0130.6460.8230.8790.8730.8530.8010.8920.8800.897
NMI0.2700.2810.4980.6370.5990.5510.4960.6300.6090.658
ARI0.0740.1750.5490.6820.6640.6120.5130.6620.6760.719
F10.0090.6400.8240.8780.8740.8540.8000.8920.8800.896
CoraACC0.0710.5850.6260.5330.6930.6420.6890.6940.7100.725
NMI0.4000.4590.4510.3880.5300.4430.5370.5310.5390.556
ARI0.1600.3180.3720.2840.4500.3660.4480.4650.4710.506
F10.0100.5390.6110.3610.6890.6490.6560.6960.7010.705
DBLPACC0.0120.4840.6080.6540.6740.5790.6430.6780.6790.689
NMI0.2520.1850.2520.3110.3620.2680.3460.3430.3110.341
ARI0.0170.1310.2560.3270.3470.2170.2800.3410.3280.356
F10.0050.4910.6040.6370.6630.5660.6480.6750.6720.688
CiteSeerACC0.0130.4480.6100.6020.6660.6360.6730.6730.6780.686
NMI0.3350.1620.3350.3260.3910.3780.4140.4120.4160.432
ARI0.1290.1560.3440.3340.3890.3490.4200.4280.4340.447
F10.0270.3850.5630.5620.5910.6020.6270.6250.6330.644
CornellACC0.0510.4870.3820.4510.5080.4150.5140.4920.4680.517
NMI0.1270.1390.0650.1420.1420.0640.1700.1400.1420.154
ARI0.0020.1410.0460.1220.1130.0600.1540.1420.0740.156
F10.0180.3230.2480.3130.3100.0380.3160.3310.3520.367
TexasACC0.1970.5030.4440.5300.5190.4550.5090.5110.5080.546
NMI0.0740.0730.0690.0560.1200.0800.0390.0820.1010.106
ARI0.0050.0130.0610.0750.0910.0440.0270.0760.1030.111
F10.0450.2330.2550.2050.2920.2430.2000.2530.2700.294
WisconsinACC0.0920.4010.4870.5340.5150.4280.6160.5690.5500.611
NMI0.0920.0470.0820.1400.1330.1390.2410.2180.1740.262
ARI0.0030.0600.0800.1920.1590.1070.2330.2240.1630.256
F10.0280.2860.3130.2820.2500.3520.3340.3580.3440.374
), ArticleFig(id=1325493663337423419, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=表2, caption=

MMCC和基准算法的指标值对比结果

, figureFileSmall=null, figureFileBig=null, tableContent=
数据集指标EdMotmotif2vecGAESDCNCaEGCNDAEGCAGCNMPTGCMMCC-3
(本文)
MMCC-4
(本文)
注:黑体数值表示最优结果,下划线数值表示次优结果。
ACMACC0.0130.6460.8230.8790.8730.8530.8010.8920.8800.897
NMI0.2700.2810.4980.6370.5990.5510.4960.6300.6090.658
ARI0.0740.1750.5490.6820.6640.6120.5130.6620.6760.719
F10.0090.6400.8240.8780.8740.8540.8000.8920.8800.896
CoraACC0.0710.5850.6260.5330.6930.6420.6890.6940.7100.725
NMI0.4000.4590.4510.3880.5300.4430.5370.5310.5390.556
ARI0.1600.3180.3720.2840.4500.3660.4480.4650.4710.506
F10.0100.5390.6110.3610.6890.6490.6560.6960.7010.705
DBLPACC0.0120.4840.6080.6540.6740.5790.6430.6780.6790.689
NMI0.2520.1850.2520.3110.3620.2680.3460.3430.3110.341
ARI0.0170.1310.2560.3270.3470.2170.2800.3410.3280.356
F10.0050.4910.6040.6370.6630.5660.6480.6750.6720.688
CiteSeerACC0.0130.4480.6100.6020.6660.6360.6730.6730.6780.686
NMI0.3350.1620.3350.3260.3910.3780.4140.4120.4160.432
ARI0.1290.1560.3440.3340.3890.3490.4200.4280.4340.447
F10.0270.3850.5630.5620.5910.6020.6270.6250.6330.644
CornellACC0.0510.4870.3820.4510.5080.4150.5140.4920.4680.517
NMI0.1270.1390.0650.1420.1420.0640.1700.1400.1420.154
ARI0.0020.1410.0460.1220.1130.0600.1540.1420.0740.156
F10.0180.3230.2480.3130.3100.0380.3160.3310.3520.367
TexasACC0.1970.5030.4440.5300.5190.4550.5090.5110.5080.546
NMI0.0740.0730.0690.0560.1200.0800.0390.0820.1010.106
ARI0.0050.0130.0610.0750.0910.0440.0270.0760.1030.111
F10.0450.2330.2550.2050.2920.2430.2000.2530.2700.294
WisconsinACC0.0920.4010.4870.5340.5150.4280.6160.5690.5500.611
NMI0.0920.0470.0820.1400.1330.1390.2410.2180.1740.262
ARI0.0030.0600.0800.1920.1590.1070.2330.2240.1630.256
F10.0280.2860.3130.2820.2500.3520.3340.3580.3440.374
), ArticleFig(id=1325493663756853820, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=EN, label=Table 3, caption=Configurations of Model Variants in Ablation Study, figureFileSmall=null, figureFileBig=null, tableContent=
模型${\mathcal{M}_3}$${\mathcal{M}_4}$多视图融合跨视图对比自监督聚类
OUR1×××××
OUR2√××××
OUR3√×√××
OUR4√×√√×
OUR5×√×××
OUR6×√√××
OUR7×√√√×
MMCC-3√×√√√
MMCC-4×√√√√
), ArticleFig(id=1325493664184672829, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, language=CN, label=表3, caption=

消融实验中模型变体的配置

, figureFileSmall=null, figureFileBig=null, tableContent=
模型${\mathcal{M}_3}$${\mathcal{M}_4}$多视图融合跨视图对比自监督聚类
OUR1×××××
OUR2√××××
OUR3√×√××
OUR4√×√√×
OUR5×√×××
OUR6×√√××
OUR7×√√√×
MMCC-3√×√√√
MMCC-4×√√√√
)], attaches=null, journal=Journal(id=1324759606947315908, delFlag=0, nameCn=计算机研究与发展, nameEn=Journal of Computer Research and Development, nameHistory1=null, nameHistory2=null, issn=1000-1239, eissn=null, cn=11-1777/TP, coden=null, periodic=0, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=k2l8zgqwstuxxevFX2qxVA==, journalPrice=null, startedYear=null, abbrevIsoEn=Journal of Computer Research and Development, journalRemark=null, publicationField=null, createdTime=1793776903141, updatedTime=1793850698612, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=J, firstLetterEn=J, subjectCode=Electronic Information, subjectName=null, subjectCodeEn=Electronic Information, subjectNameEn=null, picCn=k2l8zgqwstuxxevFX2qxVA==, picEn=r6k0Ry+RD+c/VXj5QarmDw==, jcr=null, cjcr=null, exts=[JournalExt(id=1325069127710286688, language=CN, name=计算机研究与发展, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1793850698640, updatedTime=1793850698640, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://crad.ict.ac.cn/journalx/authorLogOn.action, submissionEditorUrl=https://crad.ict.ac.cn/journalx/editorLogOn.action, submissionReviewUrl=https://crad.ict.ac.cn/journalx/expertLogOn.action, submissionCeEditorUrl=https://crad.ict.ac.cn/journalx/editorInChiefLogOn.action, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1325069127777395553, language=EN, name=Journal of Computer Research and Development, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1793850698656, updatedTime=1793850698656, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://crad.ict.ac.cn/journalx/authorLogOn.action, submissionEditorUrl=https://crad.ict.ac.cn/journalx/editorLogOn.action, submissionReviewUrl=https://crad.ict.ac.cn/journalx/expertLogOn.action, submissionCeEditorUrl=https://crad.ict.ac.cn/journalx/editorInChiefLogOn.action, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1324759914146529478, websiteList=[Website(id=1325070300139892771, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1324759914146529478, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN, language=CN, createTime=1793850978169, createBy=18614031015, updateTime=1793851040259, updateBy=18614031015, name=计算机研究与发展-中文, tplId=1146099689490845704, title=计算机研究与发展, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1325070893751349360, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=articleTextType, value=kx, createTime=1793851119697, updateTime=1793851119697, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893730377837, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=banner, value=null, createTime=1793851119692, updateTime=1793851119692, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893772320883, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=grayFlag, value=0, createTime=1793851119702, updateTime=1793851119702, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893721989228, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=logo, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/file/pic?fileId=U27tDsy717YF7XSPA+Vwdw==, createTime=1793851119690, updateTime=1793851119690, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893784903797, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=minRunFlag, value=0, createTime=1793851119705, updateTime=1793851119705, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893742960751, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/file/pic, createTime=1793851119695, updateTime=1793851119695, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893780709492, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=silenceFlag, value=0, createTime=1793851119704, updateTime=1793851119704, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893738766446, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1793851119694, updateTime=1793851119694, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893755543665, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=themeColor, value=null, createTime=1793851119698, updateTime=1793851119698, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893763932274, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=themeStyle, value=null, createTime=1793851119700, updateTime=1793851119700, creator=18614031015, updator=18614031015)]), Website(id=1325070300580294698, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1324759914146529478, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN, language=EN, createTime=1793850978274, createBy=18614031015, updateTime=1793851036638, updateBy=18614031015, name=计算机研究与发展-英文, tplId=1146101810881728533, title=Journal of Computer Research and Development, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1325070925288321148, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=articleTextType, value=kx, createTime=1793851127216, updateTime=1793851127216, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925263155321, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=banner, value=null, createTime=1793851127210, updateTime=1793851127210, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925309292671, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=grayFlag, value=0, createTime=1793851127221, updateTime=1793851127221, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925254766712, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=logo, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/file/pic?fileId=U27tDsy717YF7XSPA+Vwdw==, createTime=1793851127208, updateTime=1793851127208, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925326069889, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=minRunFlag, value=0, createTime=1793851127225, updateTime=1793851127225, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925279932539, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/file/pic, createTime=1793851127214, updateTime=1793851127214, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925317681280, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=silenceFlag, value=0, createTime=1793851127223, updateTime=1793851127223, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925271543930, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1793851127212, updateTime=1793851127212, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925292515453, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=themeColor, value=null, createTime=1793851127217, updateTime=1793851127217, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925300904062, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=themeStyle, value=null, createTime=1793851127219, updateTime=1793851127219, creator=18614031015, updator=18614031015)])], journalTitle=计算机研究与发展, weixinUrl=null, journalUrl=https://crad.ict.ac.cn/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Journal of Computer Research and Development, journalPhotoCn=k2l8zgqwstuxxevFX2qxVA==, journalPhotoEn=r6k0Ry+RD+c/VXj5QarmDw==, journalFirstLetter=J, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false, interPubPlatform=, interPubPlatformUrl=null), detailUrlCn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/10.7544/issn1000-1239.202550356, detailUrlEn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/10.7544/issn1000-1239.202550356, pdfUrlCn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/PDF/10.7544/issn1000-1239.202550356, pdfUrlEn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/PDF/10.7544/issn1000-1239.202550356, aliStartDate=0, aliEndDate=0, collectionFlag=false, citedCount=null, citedUrl=null, previewStatus=0, delFlag=0, hasFullText=1, orderTime=1769875200000, fullTextJson=null, articleText=null, reference=null)
收藏切换
模体感知的多视图协同聚类优化算法
刘美麟 1 , 李华 1 , 郑文萍 1, 2, 3, *
计算机研究与发展 | 人工智能 2026,63(2): 324-337
收起
收藏切换
计算机研究与发展 |人工智能 2026 , 63 (2) : 324 -337
模体感知的多视图协同聚类优化算法
全屏
[Author(id=1325493623881605634, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=834754720@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1325493624401699332, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493623881605634, language=EN, stringName=Meilin Liu, firstName=Meilin, middleName=null, lastName=Liu, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006, bio={"content":"

born in 1996. PhD candidate. Student member of CCF. Her main research interest includes complex network analysis

"}, bioImg=null, bioContent=

born in 1996. PhD candidate. Student member of CCF. Her main research interest includes complex network analysis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1325493626448519685, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493623881605634, language=CN, stringName=刘美麟, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1山西大学计算机与信息技术学院 太原 030006, bio={"img":"l7LqkR1upwg0T4salU033A==","content":"

1996年生。博士研究生。CCF学生会员。主要研究方向为复杂网络分析

"}, bioImg=l7LqkR1upwg0T4salU033A==, bioContent=

1996年生。博士研究生。CCF学生会员。主要研究方向为复杂网络分析

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1325493620219978232, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=1, ext=[AuthorCompanyExt(id=1325493620526162425, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493620555522554, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1山西大学计算机与信息技术学院 太原 030006)])]), Author(id=1325493626821812743, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1325493628977685001, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493626821812743, language=EN, stringName=Hua Li, firstName=Hua, middleName=null, lastName=Li, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006, bio={"content":"

born in 2001. Master candidate. Student member of CCF. Her main research interest includes complex network analysis

"}, bioImg=null, bioContent=

born in 2001. Master candidate. Student member of CCF. Her main research interest includes complex network analysis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1325493629350978058, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493626821812743, language=CN, stringName=李华, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1山西大学计算机与信息技术学院 太原 030006, bio={"img":"tKZXn05ELG0b6JMTEuEAJQ==","content":"

2001年生。硕士研究生。CCF学生会员。主要研究方向为复杂网络分析

"}, bioImg=tKZXn05ELG0b6JMTEuEAJQ==, bioContent=

2001年生。硕士研究生。CCF学生会员。主要研究方向为复杂网络分析

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1325493620219978232, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=1, ext=[AuthorCompanyExt(id=1325493620526162425, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493620555522554, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1山西大学计算机与信息技术学院 太原 030006)])]), Author(id=1325493631024505356, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=wpzheng@sxu.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1325493632689644048, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493631024505356, language=EN, stringName=Wenping Zheng, firstName=Wenping, middleName=null, lastName=Zheng, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, *, address=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006
2Key Laboratory of Computation Intelligence and Chinese Information Processing (Shanxi University), Ministry of Education, Taiyuan 030006
3Institute of Intelligent Information Processing, Shanxi University, Taiyuan 030006, bio={"content":"

born in 1979. PhD, professor, PhD supervisor. Member of CCF. Her main research interests include complex network analysis and bioinformatics

"}, bioImg=null, bioContent=

born in 1979. PhD, professor, PhD supervisor. Member of CCF. Her main research interests include complex network analysis and bioinformatics

, aboutCorrespAuthor=null), CN=AuthorExt(id=1325493633134240273, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, authorId=1325493631024505356, language=CN, stringName=郑文萍, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, *, address=1山西大学计算机与信息技术学院 太原 030006
2计算智能与中文信息处理教育部重点实验室(山西大学) 太原 030006
3山西大学智能信息处理研究所 太原 030006, bio={"img":"x2PFQ7SpDLkSYh1vM+vbIg==","content":"

1979年生。博士,教授,博士生导师。CCF会员。主要研究方向为复杂网络分析、生物信息学

"}, bioImg=x2PFQ7SpDLkSYh1vM+vbIg==, bioContent=

1979年生。博士,教授,博士生导师。CCF会员。主要研究方向为复杂网络分析、生物信息学

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1325493620219978232, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=1, ext=[AuthorCompanyExt(id=1325493620526162425, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Computer and Information Technology, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493620555522554, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493620219978232, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1山西大学计算机与信息技术学院 太原 030006)]), AuthorCompany(id=1325493623067910651, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=2, ext=[AuthorCompanyExt(id=1325493623076299260, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623067910651, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Key Laboratory of Computation Intelligence and Chinese Information Processing (Shanxi University), Ministry of Education, Taiyuan 030006), AuthorCompanyExt(id=1325493623084687869, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623067910651, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2计算智能与中文信息处理教育部重点实验室(山西大学) 太原 030006)]), AuthorCompany(id=1325493623168573950, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, xref=3, ext=[AuthorCompanyExt(id=1325493623176962559, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623168573950, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3Institute of Intelligent Information Processing, Shanxi University, Taiyuan 030006), AuthorCompanyExt(id=1325493623185351168, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1325493590348145109, companyId=1325493623168573950, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3山西大学智能信息处理研究所 太原 030006)])])]
刘美麟1 , 李华1, 郑文萍1, 2, 3, *
作者信息
  • 1山西大学计算机与信息技术学院 太原 030006
  • 2计算智能与中文信息处理教育部重点实验室(山西大学) 太原 030006
  • 3山西大学智能信息处理研究所 太原 030006
通讯作者:
郑文萍()
作者简介:

1996年生。博士研究生。CCF学生会员。主要研究方向为复杂网络分析

2001年生。硕士研究生。CCF学生会员。主要研究方向为复杂网络分析

1979年生。博士,教授,博士生导师。CCF会员。主要研究方向为复杂网络分析、生物信息学

Motif-Aware Multi-View Cooperative Clustering Optimization Algorithm
Meilin Liu1 , Hua Li1, Wenping Zheng1, 2, 3, *
Affiliations
  • 1School of Computer and Information Technology, Shanxi University, Taiyuan 030006
  • 2Key Laboratory of Computation Intelligence and Chinese Information Processing (Shanxi University), Ministry of Education, Taiyuan 030006
  • 3Institute of Intelligent Information Processing, Shanxi University, Taiyuan 030006
  • born in 1996. PhD candidate. Student member of CCF. Her main research interest includes complex network analysis

    born in 2001. Master candidate. Student member of CCF. Her main research interest includes complex network analysis

    born in 1979. PhD, professor, PhD supervisor. Member of CCF. Her main research interests include complex network analysis and bioinformatics

出版时间: 2026-02-01 doi: 10.7544/issn1000-1239.202550356
文章导航
收藏切换

图神经网络通过迭代聚合邻域特征学习图的嵌入表示,已广泛应用于图数据分析。现有方法主要关注低阶点边交互,而对以模体为载体的高阶成组交互模式关注不足,导致复杂网络中的高阶依赖关系难以被充分捕捉。模体作为网络中频繁出现的功能性子结构,能够有效揭示节点间的高阶语义关联,而模体共现视图则为刻画此类交互模式提供了新的表征视角。然而,模体共现视图的弱连通性限制了图神经网络的消息传递能力,影响全局信息的有效传播。针对此提出模体感知的多视图协同聚类优化算法(motif-aware multi-view cooperative clustering optimization algorithm,MMCC),通过自适应多视图融合机制充分挖掘高阶拓扑信息,同时利用对比学习增强不同视图间的表征一致性,从而缓解消息传递受限问题。具体而言,首先MMCC基于不同模体构建多个模体共现视图,并设计基于语义注意力的多视图自编码器动态学习不同模体视图的重要性,实现各视图的自适应融合;其次,引入对比学习约束原始视图与模体共现视图的嵌入空间一致性,缓解因模体共现视图弱连通性导致的消息传递受限问题;最后,通过优化基于KL散度的目标函数,实现特征学习与聚类任务的联合优化。在7个真实网络数据集上的聚类结果表明,MMCC在准确率(accuracy,ACC)、标准化互信息(normalized mutual information,NMI)、F1分数(F1)和调整兰德系数(adjusted Rand index,ARI)上较8个基准算法展现更显著的优势。

复杂网络  /  模体  /  高阶相互作用  /  图神经网络  /  对比学习  /  注意力机制

Graph neural networks (GNNs) learn graph embeddings by iteratively aggregating neighborhood features and have been widely applied in graph data analysis. Existing methods primarily focus on low-order node-edge interactions while overlooking high-order group interaction patterns captured by motifs. This limitation hinders the effective modeling of high-order dependencies in complex networks. Motifs, as frequently occurring functional substructures in networks, can effectively reveal high-order semantic relationships among nodes. The motif co-occurrence view provides a novel perspective for representing such interactions. However, the weak connectivity of motif co-occurrence views restricts message passing in GNNs, thereby impeding the effective propagation of global information. To address this issue, we propose a motif-aware multi-view cooperative clustering optimization algorithm (MMCC), which fully exploits high-order topological information through an adaptive multi-view fusion mechanism while enhancing representation consistency across different views via contrastive learning to alleviate the issue of restricted message passing. Specifically, MMCC first constructs multiple motif co-occurrence views based on different motif structures and employs a semantic attention-based multi-view auto-encoder to dynamically learn the importance of different motif views, achieving adaptive fusion. Then, contrastive learning is introduced to enforce embedding space consistency between the original view and motif co-occurrence views, mitigating the message passing limitations caused by weak connectivity. Finally, by optimizing a KL-divergence-based objective function, MMCC jointly optimizes feature learning and clustering tasks. Experimental results on seven real-world networks demonstrate that MMCC outperforms eight baseline algorithms in clustering accuracy (ACC), normalized mutual information (NMI), F1-score (F1), and adjusted Rand index (ARI), highlighting its effectiveness in high-order network clustering.

complex network  /  motif  /  high-order interaction  /  graph neural network  /  contrastive learning  /  attention mechanism
刘美麟, 李华, 郑文萍. 模体感知的多视图协同聚类优化算法. 计算机研究与发展, 2026 , 63 (2) : 324 -337 . DOI: 10.7544/issn1000-1239.202550356
Meilin Liu, Hua Li, Wenping Zheng. Motif-Aware Multi-View Cooperative Clustering Optimization Algorithm[J]. Journal of Computer Research and Development, 2026 , 63 (2) : 324 -337 . DOI: 10.7544/issn1000-1239.202550356
图聚类作为揭示复杂网络中潜在簇结构的核心任务,对深入理解网络的拓扑结构、功能机制和演化行为具有重要意义[1-2]。现有的深度图聚类方法将节点映射至低维嵌入空间并结合聚类算法进行划分,已展现出良好的性能[3-5]。然而,这类方法通常侧重于点边层面的低阶交互建模,难以有效捕捉复杂网络中普遍存在的高阶结构依赖,限制了对节点间潜在语义关联的深入挖掘,进而影响聚类效果。因此,设计具备高阶结构感知能力的图聚类模型,已成为当前图学习领域的重要研究方向。
Milo等人[6]正式提出“模体(motif)”的概念,拓展了传统网络结构分析范式,将关注点从二元边关系扩展至多节点间的高阶交互模式。作为网络中具有语义指示性的基本结构单元,模体在建模高阶交互关系方面展现出独特优势[7-11]。基于模体构建的模体共现视图突破了传统图结构的边连接限制,为捕捉复杂高阶依赖关系提供了新的视角[12-13]。然而,网络中往往同时包含多种类型的模体,不同模体在语义层次和功能属性上存在显著差异[14]。单一模体视图难以全面刻画网络中多样化的高阶结构特征,限制了聚类性能的进一步提升。
因此,如何在多模体共现视图之间实现有效协同,充分发挥各类模体所蕴含的语义信息,成为实现高质量图聚类的关键挑战。一方面,不同模体视图对网络的表示能力和聚类贡献存在差异,亟需设计自适应融合机制以动态评估各视图的重要性。另一方面,由于模体在真实网络中的分布通常较为稀疏[6],基于其构建的视图可能存在多个弱连通分支,导致传统图神经网络(graph neural network,GNN)[15]在这些视图上的信息传播受限,难以捕捉有效的上下文信息。此外,不同模体视图之间还存在显著的表示不一致性,进一步加剧了高阶信息的融合难度。
针对上述问题,本文提出模体感知的多视图协同聚类优化算法(motif-aware multi-view cooperative clustering optimization algorithm,MMCC),通过自适应多视图融合机制动态学习不同视图的重要性,并利用对比学习增强不同视图的表征一致性,有效缓解模体共现视图弱连通性带来的消息传递受限问题。MMCC通过构建多样化的模体共现视图揭示网络的高阶聚类特性,设计基于语义注意力的图自编码器以捕获各视图对表示学习的不同贡献;同时,引入对比学习策略改善模体共现视图弱连通导致的消息传递受限问题;最后,通过引入基于KL散度的目标函数,实现了表示学习与聚类过程的端到端联合优化。在7个真实网络上进行图聚类任务,并与8个基准算法进行比较,所提算法MMCC在准确率(accuracy,ACC)、标准化互信息(normalized mutual information,NMI)、F1分数(F1)和调整兰德系数(adjusted Rand index,ARI)方面表现良好。
本节回顾了与图聚类任务相关的研究,主要包括深度图聚类方法与模体感知的图聚类方法。
深度图聚类方法通过学习节点的低维嵌入表示,并结合聚类算法实现节点划分,近年来成为主流研究趋势。图卷积网络(graph convolutional network,GCN)[5]基于谱域卷积实现邻域信息聚合,为其发展奠定了理论基础。随后,GAT(graph attention network)[16]引入注意力机制,使模型能够动态识别并聚焦于关键邻居节点,从而提升了嵌入表示的鲁棒性。
为进一步挖掘图中丰富的结构信息,多视图对比学习被引入图聚类中。GraphCL(graph contrastive learning)[17]构建了多种视图,通过最大化同一节点在不同视图下嵌入表示的一致性提升表征的鲁棒性。FMDC(fast multi-view discrete clustering)[18]基于锚点构造多视图图结构,结合离散聚类目标,在保持效率的同时实现多视角结构信息的协同建模。
将图结构与自编码器(graph autoencoder,GAE)框架结合进一步提升了聚类性能。如图自编码器[19]将自监督重构思想引入图表征学习,通过邻接矩阵重建提升结构感知能力。DAEGC(deep attentional embedded graph clustering)[20]将图自编码器与注意力机制结合,通过重构邻接矩阵实现结构约束的嵌入学习,并提升聚类性能。SDCN(structural deep clustering network)[21]融合图神经网络与深度自编码器,联合学习结构与属性表示以优化聚类效果。CaEGCN(cross-attention fusion based enhanced graph convolutional network)[22]以交叉注意力连接图自编码器与内容自编码器,增强结构与语义的协同建模能力,提升嵌入表达的判别性。
掩码机制也被引入图自编码器的框架中以增强对局部结构与属性的建模能力。GraphMAE(masked graph autoencoder)[14]采用节点属性掩码引导模型重建缺失特征,提升无监督图表征的泛化能力。MaskGAE(masked graph autoencoder)[23]通过引入拓扑结构的掩码机制,并结合结构预测任务,有效提升了模型对邻接关系的建模能力。HC-GAE(hierarchical cluster-based GAE)[24]引入基于硬/软节点分配的图层次压缩与重建机制,提取双向层次结构特征,有效缓解传统图自编码器中的过平滑问题,提升了节点与图级任务的表示效果。
尽管上述方法取得了良好效果,但其建模主要停留在点-边层面,难以捕捉模体等高阶结构中所蕴含的复杂群体交互模式,限制了对网络高阶语义关系的建模能力。针对这一问题,亟需引入具备高阶结构感知能力的建模机制。
模体作为网络中的高阶结构单元,在揭示节点之间的潜在群体交互模式方面具有重要意义。gSparsify(graph sparsification)[11]基于模体在边中的贡献度评估,对图结构进行稀疏化,保留模体内的边以提升聚类性能。MAPPR(motif-based approximate personalized pagerank)[25]利用模体电导设计个性化PageRank算法,实现对局部社群结构的高效发现。EdMot(edge enhancement approach for motif-aware community detection)[12]在模体增强图中重构边权,并结合社区检测算法挖掘潜在社区结构。HM-Modularity(harmonic motif modularity)[26]定义模体调和模块度指标以评估划分质量,辅助社区发现。然而,这些方法依赖人工设定的结构指标与规则,缺乏可学习的特征提取机制,难以充分挖掘模体信息在不同上下文中的泛化潜力。
近年来,模体感知机制逐渐融入图表示学习框架中,以增强模型对高阶依赖结构的建模能力。RUM(network representation learning using motifs)[27]通过构建模体辅助网络引导随机游走,提升节点嵌入的表达能力。motif2vec[28]在异质网络中构建模体邻接矩阵,通过有偏游走策略实现高阶结构感知的表示学习。MODEL(motif-based deep feature learning algorithm)[29]结合模体子图采样与对比学习,优化模体内部节点表示的一致性。LEMON(learning embeddings by leveraging motifs of networks)[30]构建了由模体超顶点与原始顶点组成的异质网络,利用异质游走策略融合局部与全局结构特征。此类方法通过将模体检测与游走策略相结合,在保持结构连通性的同时增强了对高阶相似性的建模能力。
在图神经网络中集成模体结构的研究也逐渐展开。HONE(higher-order network embeddings)[31]提出通用高阶结构嵌入框架,通过最小化Bregman散度优化模体矩阵,实现了结构感知的聚类目标。Motif-GAT[32]基于多跳模体邻接矩阵设计注意力机制,扩展感受野并提升节点对高阶结构的感知能力。MPTGC(motif-based prompt tuning for graph clustering)[33]设计基于模体的提示机制以缓解长尾噪声影响,并优化表示学习过程。然而,这些方法忽视了不同模体间的一致性与差异性,同时模体共现视图通常呈现弱连通性,限制了GCNs的消息传递能力。
本文所提出的MMCC算法包括3个关键模块,其中,多视图融合模块动态学习不同模体视图的重要性权重,实现自适应加权融合;跨视图对比模块利用对比学习约束原始图与模体共现网络的表征一致性以缓解消息传递受限问题;自监督聚类模块通过KL散度优化聚类分布以实现端到端的节点划分。
给定一个图$ G = (V,E,{\boldsymbol{X}}) $,其中$ V = \{ {v_1},{v_2},…,{v_{|V|}}\} $为节点集,$ E = \{ {e_1},{e_2},…,{e_{|E|}}\} $为边集,$ {\boldsymbol{X}} \in {\mathbb{R}^{n \times f}} $为节点的特征矩阵,其中f为节点的特征维度。对于图G中的任意节点${v_i}$,若存在一条边将其与节点${v_j}$相连,则称${v_j}$是${v_i}$的直接邻居,节点${v_i}$的邻居节点集为$N({v_i}) = \{ {v_j}| ({v_i},{v_j}) \in E \wedge {v_j} \in V\} $,$ d({v_i}) = |N({v_i})| $为该节点的度。图G的邻接矩阵记为$ {\boldsymbol{A}} \in {\mathbb{R}^{n \times n}} $,若节点${v_i}$和节点${v_j}$之间存在边,则${a_{ij}} = 1$,否则${a_{ij}} = 0$。图G的度矩阵为$ {\boldsymbol{D}} \in {\mathbb{R}^{n \times n}} $,其中$ {d_{ii}} = d({v_i}) = \displaystyle\sum_j {{a_{ij}}} $。
对于图$G' = (V',E')$,若$V' \subseteq V$且$E' \subseteq E$,称$G'$是G的一个子图,记作$ G' \subseteq G $,对于子图$G'$中的任意节点对${v_i},{v_j} \in V'$,若存在边${e_{ij}} = ({v_i},{v_j}) \in E$且边${e_{ij}} \in E'$,则称$G'$为G相对于节点子集$V'$的导出子图,记作$ G[V'] $。
在图G中,若节点vi和节点vj之间存在路径,则称节点vi和节点vj是连通的。若图G中任意2个节点vi,vj之间都存在路径,则称图G是连通的,否则图G不连通。连通图G的点割集是一个节点集$S \subseteq V $,使得$ G - S $不连通。图G的连通度为最小点割集的节点数,即使得$ G - S $不连通的S的最小规模,记作k(G)。若k(G)=0,则图G不连通;若k(G)≥k,则称图G是k-连通图。
网络模体$M = ({V_M},{E_M})$是一个小的连通子图,其规模$s = \left| {{V_M}} \right|$。具有相同节点数的模体可以表现为多种不同的子图结构,反映了不同网络的拓扑特性,是复杂网络的基本构建块。令${\mathcal{M}_s} = \{ {M_{s.1}},{M_{s.2}},…,{M_{s.k}}\} $表示规模为s的k个不同构模体的集合,图1展示了2-模体集${\mathcal{M}_2} = \{ {M_2}\} $、3-模体集${\mathcal{M}_3} = \{ {M_{3.1}},{M_{3.2}}\} $和4-模体集${\mathcal{M}_4} = \{ {M_{4.1}},{M_{4.2}},{M_{4.3}},{M_{4.4}},{M_{4.5}},{M_{4.6}}\} $。若某个子图$ G[V'] $与模体M同构,则认为节点子集$ V' $是模体M的1个实例,模体M的所有实例集记为$I[M]$,包含节点u的模体实例集被记为${I_u}[M]$。
对于模体M,当且仅当M中节点的连通度不小于k,称M为k-连通模体。在k-连通模体中,每对节点之间至少存在k条不同的点不相交路径。图1中M2,M3.1,M4.1,M4.3,M4.4为单连通模体,M3.2,M4.2,M4.5,M4.6为二连通模体。在单连通模体中,若去掉任意一条边,模体可能会不连通;而在二连通模体中,即使去掉任意一条边,模体依然保持连通性。二连通模体通常比单连通模体具有更强的连通性。
MMCC旨在通过自适应融合机制动态学习不同模体共现视图的重要性,实现各视图的自适应融合。同时,借助对比学习增强不同视图间的表征一致性,以有效缓解因弱连通性导致的消息传递受限问题。该算法主要包括多视图融合、跨视图对比和自监督聚类3个模块,其整体框架如图2所示。
在真实网络中,节点间的交互不仅局限于低阶的点边交互,还包含丰富的高阶成组交互模式,如三角形、团簇、星形结构等,这些模式可以通过模体进行建模。不同类型的模体在网络中承载着不同的语义信息,对网络表示学习和聚类任务的贡献也存在差异。然而,单一模体共现视图难以全面刻画网络中多样化的高阶拓扑关系。为此,本文提出基于多模体共现视图的语义注意力融合机制,以自适应方式整合不同尺度的高阶结构特征,从而增强网络表征能力,为后续的聚类任务提供更稳定和全面的节点表示。
给定原始网络$ G = (V,E,{\boldsymbol{X}}) $,V表示节点集,E表示边集,X表示节点的特征矩阵。本文选取k种不同的模体结构(如3-模体(k=2),4-模体(k=6)),构建k个模体共现视图$ {G_1},{G_2}, … ,{G_k} $。对于每种模体Mk通过统计节点对在不同模体中的共现关系来定义模体共现矩阵${{\boldsymbol{C}}_k} \in {\mathbb{R}^{n \times n}}$,${{\boldsymbol{C}}_k} $的元素$c_{ij} $定义为:
$ {c_{ij}} = \displaystyle\sum_{m \in I[{M_k}]} {{\mathbb{I}}(\{ {v_i},{v_j}\} \subset m)} \;,$
其中$I[M_k] $为与模体$M_k $同构的模体实例集,m为其中的某个模体实例,${\mathbb{I}}( \cdot )$为指示函数,当节点vi与节点vj共现于模体实例m时,${\mathbb{I}}( \cdot ) = 1$,否则${\mathbb{I}}( \cdot ) = 0$。
基于此,构建模体共现视图Gk,其中节点是原始网络中的节点集,边集由模体共现关系生成,边的权重为cij,即节点vi和节点vj之间关于模体Mk的共现频率。图3展示了基于模体M4.1进行模体共现视图构建的实例。图3最左侧为原始网络,利用模体M4.1识别出多个同构的模体实例集$I[M_{4.1}] $,进一步构建模体共现视图。通过模体共现视图的构建,我们能够系统地刻画节点在不同模体视角下的协同关系。这种方式不仅能缓解单一视图信息不足的问题,还能通过不同模体视角的互补性,提高最终节点表示的完整性和区分能力。
给定k个模体共现视图$ {G_1},{G_2}, … ,{G_k} $,每个视图对应的邻接矩阵为$ {{\boldsymbol{C}}_1},{{\boldsymbol{C}}_2}, … ,{{\boldsymbol{C}}_k} $,采用2层GCN计算节点在该视图下的嵌入表示:
$ {\boldsymbol{Z}}_i^{(l)} = \sigma \left(\tilde {\boldsymbol{D}}_i^{ - \tfrac{1}{2}}{\tilde {\boldsymbol{C}}_i}\tilde {\boldsymbol{D}}_i^{ - \tfrac{1}{2}}{\boldsymbol{Z}}_i^{(l - 1)}{\boldsymbol{W}}_i^{(l)}\right) \;,$
其中$ {\tilde {\boldsymbol{C}}_i} = {{\boldsymbol{C}}_i} + {\boldsymbol{I}} $为带自环的邻接矩阵,$ {\tilde {\boldsymbol{D}}_i} $为度矩阵,$ {\boldsymbol{Z}}_i^{(0)} = {\boldsymbol{X}} $,$ {\boldsymbol{W}}_i^{(l)} $是第l层GCN中的可学习权重矩阵,$ \sigma ( \cdot ) $是ReLU激活函数,用来缓解梯度消失问题。通过式(2)可以获得k个模体共现视图的节点表示矩阵$ {{\boldsymbol{Z}}_1},{{\boldsymbol{Z}}_2}, … ,{{\boldsymbol{Z}}_k} $。
为了有效整合不同模体共现视图的信息,采用语义注意力机制来自适应地学习不同模体共现视图在聚类任务中的重要性权重,计算公式为:
$ {\alpha _i} = \frac{1}{{\left| V \right|}}\displaystyle\sum_{v \in V} {{{\boldsymbol{q}}^{\rm T}}\sigma ({{\boldsymbol{W}}_{{\text{att}}}}{\boldsymbol{z}}_v^i + {\boldsymbol{b}})} \;,$
其中$ {\alpha _i} $表示第i个模体共现视图的重要性权重;$ {\boldsymbol{z}}_v^i $是节点v在第i个视图的表示向量;$ {\boldsymbol{q}} \in {\mathbb{R}^{f' \times 1}} $和$ {{\boldsymbol{W}}_{{\text{att}}}} \in {\mathbb{R}^{f' \times f}} $是权重矩阵;$ {\boldsymbol{b}} \in {\mathbb{R}^{f' \times 1}} $是偏置向量;${\boldsymbol{q}} $,${\boldsymbol{W}}_{{\mathrm{att}}} $,${\boldsymbol{b}} $都是可学习的参数;$ \sigma ( \cdot ) $是tanh激活函数,用来增强视图区分性。
之后,将所有视图对应的注意力权重$\alpha_i $经过softmax归一化得到$w_i $,以确保它们的总和为1,计算公式为:
$ {w_i} = \dfrac{{\exp ({\alpha _i})}}{{\displaystyle\displaystyle\sum_{j = 1}^k {\exp ({\alpha _j})} }} \;。$
基于学习到的注意力权重wi对各个视图的节点表示进行加权融合,得到最终的节点嵌入表示Z:
$ {\boldsymbol{Z}} = \displaystyle\sum_{i = 1}^k {{w_i}{{\boldsymbol{Z}}_i}} \;。$
该融合策略能够动态调整不同模体视图的贡献,使得最终表示能够充分整合多种高阶结构信息,为后续的聚类任务提供更优的节点表示。
在获得融合后的节点嵌入表示Z后,进一步利用解码器进行结构解码和特征解码,确保最终的节点表示能够捕捉高阶拓扑关系与关键特征。
基于内积操作计算每对节点的余弦相似性,得到重构的邻接矩阵$ {\hat {\boldsymbol{A}}} $:
$ {\hat {\boldsymbol{A}}} = \sigma ({\boldsymbol{Z}}{{\boldsymbol{W}}_{{\text{rec}}}}{{\boldsymbol{Z}}^{\rm T}}) \;,$
其中${{\boldsymbol{W}}_{{\text{rec}}}} $是可训练权重,$ \sigma ( \cdot ) $是Sigmoid激活函数,使重构后拓扑矩阵更接近0-1矩阵。该重构过程本质上是通过节点嵌入的相似性来恢复网络拓扑结构,从而保证高阶结构信息的有效传递。为了使重构的邻接矩阵$ {\hat {\boldsymbol{A}}} $尽可能接近原始邻接矩阵${\boldsymbol{A}} $,本文采用交叉熵损失函数进行优化:
$ {L_{{\text{topo}}}} = - \displaystyle\sum_{i,j} {({a_{ij}}\ln {{\hat a}_{ij}} + (1 - {a_{ij}})\ln (1 - {{\hat a}_{ij}}))} \;。$
除了结构信息,节点的特征信息同样承载了丰富的语义。为此,采用2层GCN作为解码器重构节点属性信息$ {\hat{\boldsymbol X}} $:
$ {\hat{\boldsymbol X}} = f({\boldsymbol{Z}},{\boldsymbol{A}}) = {\boldsymbol{A}}\sigma ({\boldsymbol{AZ}}{{\boldsymbol{W}}^{(0)}}){{\boldsymbol{W}}^{(1)}} \;,$
其中Z是语义融合图嵌入,${{\boldsymbol{W}}^{(0)}} $和${{\boldsymbol{W}}^{(1)}} $分别是解码器中2个卷积层的2个可训练权重矩阵,$ \sigma ( \cdot ) $是ReLU激活函数。通过这种方式,模型能够在保持局部平滑性的同时,更好地恢复节点特征信息。
为了编码模块能学习到关键的特征模式,本文最小化原始特征矩阵X与重构特征矩阵$ {\hat{\boldsymbol X}} $之间的损失:
$ {L_{{\text{feat}}}} = \frac{1}{{2\left| V \right|}}\left\| {{\boldsymbol{X}} - {\hat{\boldsymbol X}}} \right\|_{\text{F}}^2 \;,$
其中${\left\| \cdot \right\|_{\text{F}}}$表示Frobenius范数。
多视图融合模块通过模体共现视图构建、语义注意力融合以及结构和特征解码器优化,实现了对高阶结构信息的有效利用,并为后续的聚类任务提供了高质量的节点表示。
在构造模体共现视图的过程中,我们发现由高平均度模体构造的模体共现视图往往表现出较弱的连通性。如图4所示,模体M4.6相较于模体M4.1具有更高的平均度,其对应的模体共现视图中存在更多的连通分量和孤立节点。在图神经网络的消息传递过程中,弱连通图的训练依赖于有限的局部连接信息,可能会忽略全局结构及其他潜在特征,从而限制图神经网络对图结构的理解和利用。
为了缓解这一问题并确保不同视图之间表示的一致性,本文引入了对比学习机制,以强化模体共现视图和其他视图之间的结构和特征一致性,同时缓解弱连通性对消息传递的限制,使得图神经网络仍能充分学习和利用全局信息。
为了提升模型的鲁棒性,本文采用行级随机重排策略来构建特征扰动视图,并将其作为负样本视图,其公式定义为:
$ {\boldsymbol{\tilde X}} = \tau ({\boldsymbol{X}}) \;,$
其中$\tau $是一个随机置换操作,将原始特征矩阵X的行按照随机序列重新排序,得到特征扰动矩阵$ {\boldsymbol{\tilde X}} $。随后利用式(2)对原始视图以及特征扰动视图进行编码,分别得到它们的节点表示矩阵$ {{\boldsymbol{Z}}_0} $和$ \tilde {\boldsymbol{Z}} $。
在对比学习过程中,本文设定模体共现视图作为锚样本视图、特征扰动视图作为负样本视图、原始网络作为正样本视图。通过最大化正样本对的相似性并最小化负样本对的相似性,增强了节点表示的拓扑一致性,同时提升了节点表示的鲁棒性,使得模型能够更有效地捕捉节点之间的相似性和差异性。在训练过程中,采用对比学习的损失函数来优化跨视图一致性,如式(11)所示。
$ {L_{{\text{con}}}} = - \frac{1}{{2\left| V \right|}}\displaystyle\sum_{i = 1}^{\left| V \right|} \left(\frac{1}{k}\displaystyle\sum_k \left(\ln D({\boldsymbol{z}}_i^k,{\boldsymbol{z}}_i^0) + \ln (1 - D({\boldsymbol{z}}_i^k,{{\tilde {\boldsymbol{z}}}_i})) \right)\right) \;,$
其中k为模体共现视图的个数,$ {\boldsymbol{z}}_i^k $表示第k个模体共现视图中节点vi的嵌入表示,$ {\boldsymbol{z}}_i^0 $和$ {\tilde {\boldsymbol{z}}_i} $分别表示原始网络和特征扰动视图中节点vi的嵌入表示,$D( \cdot )$为判别器,用于衡量不同视图之间的相似性。
通过对比学习,本文增强了模体共现视图与原始网络视图之间的表示一致性,有效缓解了因模体共现视图弱连通性带来的信息传播障碍。此外,对比学习不仅提升了不同视图下节点表示的判别能力和鲁棒性,还强化了模型对全局结构信息的理解与建模,使得节点表示在图聚类任务中更加稳定可靠,从而进一步提升了聚类效果。
为了实现端到端的节点聚类,本文引入了自监督聚类机制,通过联合优化表征学习与聚类分配,有效提升聚类质量。具体而言,通过最小化P分布和Q分布之间的KL散度以对齐图表征与聚类分布,损失函数定义为:
$ {L_{{\text{clu}}}} = KL({\boldsymbol{P}}||{\boldsymbol{Q}}) = \displaystyle\sum_i {\displaystyle\sum_u {{{\boldsymbol{p}}_{iu}}\ln \frac{{{{\boldsymbol{p}}_{iu}}}}{{{{\boldsymbol{q}}_{iu}}}}} } \;,$
其中P是辅助目标分布,${\boldsymbol{p}}_{iu} $为样本vi属于类簇u的期望概率;Q是软标签分布,${\boldsymbol{q}}_{iu} $为样本vi属于类簇u的概率。首先,在嵌入Z上利用k-means得到U个初始聚类中心${{\boldsymbol{\mu }}_1},{{\boldsymbol{\mu }}_2}, … ,{{\boldsymbol{\mu }}_U}$,再通过t-分布度量节点嵌入zi与聚类中心${{\boldsymbol{\mu }}_u}$:
$ {{\boldsymbol{q}}_{iu}} = \dfrac{{{{(1 + {{\left\| {{{\boldsymbol{z}}_i} - {{\boldsymbol{\mu }}_u}} \right\|}^2})}^{ - 1}}}}{{\displaystyle\displaystyle\sum_{u'} {{{(1 + {{\left\| {{{\boldsymbol{z}}_i} - {{\boldsymbol{\mu }}_{u'}}} \right\|}^2})}^{ - 1}}} }} \;。$
辅助目标分布P用来引导形成清晰的聚类结构,增强聚类结果的鲁棒性,piu计算过程如式(14)所示。
$ {{\boldsymbol{p}}_{iu}} = \dfrac{{{\boldsymbol{q}}_{iu}^2/\displaystyle\displaystyle\sum_i {{{\boldsymbol{q}}_{iu}}} }}{{\displaystyle\displaystyle\sum_{u'} {\left({\boldsymbol{q}}_{iu'}^2/\displaystyle\displaystyle\sum_i {{{\boldsymbol{q}}_{iu'}}} \right)} }} \;。$
最后,对每个聚类中心${{\boldsymbol{\mu }}_u}$选择对其分配概率最高的样本集合,聚类结果可以通过式(15)获得。
$ {s_u} = \mathop {\arg\,\max }\limits_u {{\boldsymbol{q}}_{iu}} \;。$
MMCC采用预训练与微调相结合的优化策略,以充分挖掘节点的属性信息和结构信息,从而提升在图聚类任务中的性能表现。在预训练阶段,MMCC通过联合优化属性重构损失、拓扑重构损失(多视图融合模块),以及对比损失(跨视图对比模块),学习具有判别性的节点嵌入表示。该阶段旨在保持嵌入空间中节点的属性一致性与结构一致性,并提升其在高阶语义下的表达能力,从而使模型具有捕获全局结构的能力。整体预训练损失定义为:
$ {L_{{\text{pre}}}} = {L_{{\text{topo}}}} + {L_{{\text{feat}}}} + {L_{{\text{con}}}}\;。$
在微调阶段,本文引入联合聚类损失函数,通过自监督聚类模块以端到端的方式对模型进行优化。该设计旨在提升节点表示的判别性与聚类适应性,从而增强模型在下游图聚类任务中的性能表现。最终的总损失函数如式(17)所示。
$ L = {L_{{\text{pre}}}} + {L_{{\text{clu}}}} \;。$
MMCC的基本框架如算法1所示。
算法1. MMCC算法。
输入:图$ G = (V,E,{\boldsymbol{X}}) $,模体集$\mathcal M_s $,聚类个数U,预训练迭代次数T1,训练迭代次数T2;
输出:节点聚类结果S。
① 根据式(1)构建规模为s的k个模体共现视图;
② 根据式(10)构建特征扰动视图;
③ for epoch in 0, 1, …, T1 then
④  根据式(2)计算模体共现视图的嵌入表示 $ {{\boldsymbol{Z}}_1},{{\boldsymbol{Z}}_2},…,{{\boldsymbol{Z}}_k} $;
⑤  根据式(5)计算语义注意力融合嵌入表示Z;
⑥  根据式(6)重构邻接矩阵${\hat{\boldsymbol A}}$;
⑦  根据式(8)重构特征矩阵$ {\hat{\boldsymbol X}} $;
⑧  最小化式(16)更新wk和模型参数;
⑨ end for
⑩ 在嵌入Z上利用k-means得到U个聚类中心${{\boldsymbol{\mu }}_u}$;
⑪ for epoch in 0, 1, …, T2 then
⑫  最小化式(17)进行模型微调更新wk和模型 参数;
⑬ end for
⑭ 根据式(15)计算最终聚类结果S;
⑮ return S。
为了评估所提出的MMCC算法的聚类性能,在4个引文网络数据集ACM[34],Cora[35],CiteSeer[36],DBLP[19]和3个Webkb数据集Texas[37],Cornell[37],Wisconsin[37]上进行实验。实验结果与EdMot,motif2vec,GAE,SDCN,CaEGCN,DAEGC,AGCN,MPTGC这8种基准算法进行了对比分析。采用ACC,NMI,F1,ARI作为评估指标衡量算法的聚类质量。各数据集的基本信息如表1所示。
在实验配置方面,为确保对比方法的公平性,所有基准算法均采用其原始论文中推荐的默认超参数设置。对于所提MMCC模型,分别基于3-模体和4-模体构建模体共现视图,记为MMCC-3和MMCC-4。在优化过程中,模型使用Adam优化器进行参数更新,学习率在引文网络数据集中设为0.001、在Webkb数据集中设为0.0001。预训练中的迭代次数设定为T1=100,聚类微调迭代次数设定为T2=50。为增强实验结果的统计可靠性,所有实验均独立重复10次,最终结果取平均值进行报告。
表2汇总了MMCC与8种基准算法在各数据集上的实验对比结果。实验结果表明,MMCC在大多数数据集上均表现优异。在ACM,Cora,CiteSeer这3个引文网络数据集上,MMCC在ACC,NMI,ARI,F1这4个评价指标上均取得了最优性能。在ACM数据集上,相较于次优算法,MMCC在ACC,NMI,ARI,F1指标上分别提升0.6%,3.3%,5.4%,0.4%;在Cora数据集上,相应提升幅度为2.1%,3.2%,7.4%,0.6%;在CiteSeer数据集上则分别提升了1.2%,3.8%,3.0%,1.7%;在DBLP数据集上,MMCC在ACC,ARI,F1上分别提升1.5%,2.6%,1.9%。此外,在Cora与CiteSeer数据集上,MMCC-3和MMCC-4均取得了最优性能。这是由于引文网络固有的同配性特征,即学术论文倾向于引用相同研究领域的文献,从而导致同类节点间形成较高的连接密度。模体共现视图能够有效增强同类节点之间的连边,并相对减少不同类别节点的连接,从而降低异类信息的干扰,使得该算法更能优化图神经网络在高同配网络中的信息传播效果,最终提升聚类性能。
此外,在Cornell,Texas,Wisconsin数据集上,MMCC同样展现了优异的性能。在Cornell数据集上,相较于性能次优的算法,MMCC在ACC,ARI,F1上分别提升了0.6%,1.3%,4.3%;在Texas数据集上,相应幅度分别提升了3.0%,7.8%,0.7%;在Wisconsin数据集上,MMCC在NMI,ARI,F1上分别提升了8.7%,9.9%,10.1%。然而需要指出的是,在这些异配性较强的数据集中,MMCC-3算法的性能表现存在一定波动,部分情况下略低于基准算法。这一现象可能归因于高异配性的网络中不同类别节点间存在大量连接,导致3-模体增强,可能意外强化异类节点间的拓扑关联,从而降低类别区分度。相比之下,4-模体由于能够捕获更复杂的高阶邻域模式,在高异配性网络中展现出更强的特征判别能力,因此获得更优的聚类效果。
为验证MMCC中多视图融合模块、跨视图对比模块以及自监督聚类模块中各子模块的有效性,我们在ACM和Cornell数据集上进行了系统的消融实验研究。如表3所示,我们设计了7种模型变体(OUR1~OUR7)和2种完整模型(MMCC-3和MMCC-4),逐步引入不同模块以评估其贡献。
图5展示了表3中各算法的实验结果。基准模型OUR1仅采用图自编码器进行节点表示,性能最为基础。引入3-模体与4-模体编码模块的OUR2和OUR5在性能上显著提升,表明高阶模体结构在捕获网络语义关联方面的重要作用。进一步加入多视图融合模块的OUR3和OUR6模型展现出更好的性能,表明该模块能够有效动态权衡不同视图的语义贡献。而整合跨视图对比模块的OUR4和OUR7模型则实现了更优的表现,验证了该模块在缓解模体共现视图弱连通性导致的消息传递受限问题方面的有效性。最终,通过引入自监督聚类模块的MMCC-3和MMCC-4模型达到了最佳性能,这一结果充分证明了在聚类任务中采用聚类损失进行模型微调的策略优势。特别值得注意的是,基于4-模体的MMCC-4相比3-模体的MMCC-3表现出更优异的性能,这是由于相较3-模体而言,利用4-模体可以从更多角度对节点的高阶作用进行建模,从而得到更具表达性的节点表示。
我们进一步分析了MMCC-3和MMCC-4在不同模体上的注意力系数分布,如图6所示。从结果可以看出,不同模体被赋予了不同的注意力系数,这表明不同模体受网络表征的影响存在差异。MMCC-3更关注二连通模体M3.2,这与社会网络分析中的复杂网络特性相对应。通常,具有三角关系的实体之间往往具有更强的同质性,这也是网络演化的基本机制之一。而MMCC-4更关注二连通模体M4.2,M4.5,M4.6,此外模体M4.4也受到了较多关注。这可能是因为M4.4具有较高的平均度,相比于其他模体更能反映网络实体间复杂的交互关系,从而增强同质性特征的学习。
此外,本文进一步统计了不同数据集上的模体显著性特征,利用基于BA无标度网络的模体丰度衡量模体的统计意义。图7展示了基准数据集上3-模体与4-模体在BA无标度网络下的丰度分布。可以发现,二连通模体M3.2,M4.5,M4.6在多个数据集上均具有较高丰度,模体M4.2与模体M4.4在部分数据集上也呈现较高丰度。这表明这些模体在真实网络中出现的频率较高,具有较强的稳定性。此外,本文还观察到引文网络数据集与Webkb网络数据集模体丰度分布存在差异,表明不同类型的网络对模体规模具有不同的依赖性。
基于MMCC关注的模体注意力系数分布和模体在BA无标度网络下的丰度分布,本文进一步研究了不同数据集模体特征间的成对相似性,结果如图8所示。从实验结果可以看出,基准数据集中模型训练过程中学习到的模体注意力系数与模体的统计丰度呈高度相关性,表明MMCC能自适应地捕获关键的拓扑模式。
此外,模体注意力机制的本质是拓扑显著性驱动,这与基于优先连接机制的网络生成过程高度契合。高频模体获得更高的注意力权重,这说明MMCC能够有效识别网络形成的内在规律,并利用模体增强表征学习,提升网络建模的表达能力。
对待网络$ G = (V,E) $,MMCC算法将模体提取作为模型训练前的预处理步骤,采用局部模体计数(local motif counting,LMC)算法[38]从图G中提取3-模体和4-模体。具体地,LMC统计每个节点的直接邻域内节点间的点不重复路径来计数3-模体,其时间复杂度为$ O(|V|{\bar d^2}) $,其中$ \bar d $为G中节点的平均度;进一步根据文献[38]中的定理1,LMC利用直接邻域中三元组信息对4-模体进行计数,时间复杂度为$ O(|V|{\bar d^3}) $。对稀疏网络而言,节点的平均度通常为$ \bar d = O(\log |V|) $,因此LMC算法是关于节点数的准线性时间复杂度$ O(|V|{\log ^3}|V|) $。由于模体提取过程在预处理阶段完成,不会增加图表示学习的代价。
为进一步评估不同模体构建策略对计算代价的影响,本节对比MMCC-3与MMCC-4在多个数据集上的预训练时间、微调时间与总训练时间,结果见图9。整体来看,MMCC-4在各阶段时间消耗较MMCC-3的平均增幅为0.26左右,主要源于MMCC-4所构建的连通4-模体共现视图数多于MMCC-3所构建的连通3-模体共现视图数。具体地,连通3-模体共有2种非同构结构,而连通4-模体具有6种非同构结构,因此MMCC-3有2个模体共现视图,而MMCC-4有6个模体共现视图。这导致MMCC-4的模型参数较MMCC-3更多,从而增加了各阶段的时间消耗。值得注意的是,随着数据规模扩大,时间消耗的增幅呈下降趋势,表明MMCC在大规模图上具有良好的可扩展性。
本节研究了隐藏层维度和学习率这2个超参数如何影响MMCC模型在ACM与Cornell数据集上的表现。
1)隐藏层维度分析。从图10可以看出,随着嵌入维度的增加,模型性能呈现先上升后下降的趋势。当嵌入维度较低时,隐空间的表征能力受限,无法充分编码多视图数据的复杂语义特征,导致关键信息在降维过程中丢失,从而影响聚类性能。当嵌入维度过高时,模型可能学习到冗余特征或噪声,同样也对模型性能产生负面影响。实验结果表明,当嵌入维度设置为32时,模型在多个评价指标上均达到最优值。
1)学习率敏感性分析。从图11可以看出,随着学习率的增加,模型性能同样呈现出先上升后下降的趋势。当学习率过小时,模型参数更新步长较小,导致优化过程收敛缓慢,甚至可能陷入局部最优,影响最终聚类效果。当学习率过大时,模型参数的更新步长过大,可能导致优化过程震荡甚至发散,使得训练损失难以稳定下降,最终降低聚类精度。具体而言,在ACM数据集上,当学习率设置为IE-3时,模型性能达到最佳状态;在Cornell数据集上,最佳学习率则为IE-4。
本文针对图聚类任务中高阶结构信息利用不足、模体共现视图弱连通性导致消息传递受限问题,提出了模体感知的多视图协同聚类优化算法(MMCC)。通过融合多样化的模体共现视图,MMCC能够有效刻画复杂网络中的高阶拓扑关系,并自适应融合不同视图的贡献。引入对比学习增强了低阶视图与高阶模体视图之间的表征一致性,缓解了因模体稀疏性导致的弱连通性问题。通过优化KL散度目标函数,实现特征学习与聚类任务的联合优化。实验结果表明,MMCC在多个真实网络数据集上相较于现有基准方法在多个指标上均表现出显著优势,验证了其在挖掘高阶语义关联和提升聚类性能方面的有效性。
未来工作可进一步设计高效的模体计数与采样算法[39-40],结合局部模体分布建模与分布式计算策略,探索适用于动态场景的模体统计方法,从而降低计算开销,并提升模型对局部高阶结构的建模能力与判别性。此外,还可研究更具判别力的模体选择机制,以增强模型对关键高阶语义的感知与表达能力。
作者贡献声明:刘美麟提出了算法思路和实验方案并撰写论文;李华进行了实验验证并修改论文;郑文萍统筹指导整个研究过程和修改论文。

参考文献 引证文献
排序方式:
1
Watts D J, Strogatz S H. Collective dynamics of ‘small-world’ networks[J]. Nature, 1998, 393(6684): 440−442
2
Girvan M, Newman M E J. Community structure in social and biological networks[J]. Proceedings of the National Academy of Science, 2002, 99(12): 7821−7826
3
Perozzi B, Al-Rfou R, Skiena S. DeepWalk: Online learning of social representations[C]//Proc of the 20th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2014: 701−710
4
Grover A, Leskovec J. node2vec: Scalable feature learning for Networks[C]//Proc of the 22nd ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2016: 855−864
5
Kipf T N, Welling M. Semi-supervised classification with graph convolutional networks[C/OL]//Proc of the 5th Int Conf on Learning Representations. Washington: ICLR, 2017[2025-06-26]. https://openreview.net/forum?id=SJU4ayYgl
6
Milo R, Shen-Orr S, Itzkovitz S, et al. Network motifs: Simple building blocks of complex networks[J]. Science, 2002, 298(5594): 824−827
7
Xia Feng, Yu Shuo, Liu Chengfei, et al. CHIEF: Clustering with higher-order motifs in big networks[J]. IEEE Transactions on Network Science and Engineering, 2022, 9(3): 990−1005
8
Benson A R, Gleich D F, Leskovec J. Higher-order organization of complex networks[J]. Science, 2016, 353(6295): 163−166
9
Wernicke S. Efficient detection of network motifs[J]. IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2006, 3(4): 347−359
10
Ribeiro P, Silva F. G-Tries: An efficient data structure for discovering network motifs[C]//Proc of the 2010 ACM Symp on Applied Computing. New York: ACM, 2010: 1559−1566
11
Zhao Peixiang. gSparsify: Graph motif based sparsification for graph clustering[C]//Proc of the 24th ACM Int Conf on Information and Knowledge Management. New York: ACM, 2015: 373−382
12
Li Peizhen, Huang Ling, Wang Changdong, et al. EdMot: An edge enhancement approach for motif-aware community detection[C]//Proc of the 25th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2019: 479−487
13
Li Pan, Dau H, Puleo G, et al. Motif clustering and overlapping clustering for social network analysis[C/OL]//Proc of the 2017 IEEE Conf on Computer Communications. Piscataway, NJ: IEEE, 2017[2025-06-26]. https://ieeexplore.ieee.org/document/8056956
14
Hou Zhenyu, Liu Xiao, Cen Yukuo, et al. GraphMAE: Self-supervised masked graph autoencoders[C]//Proc of the 28th ACM SIGKDD Conf on Knowledge Discovery and Data Mining. New York: ACM, 2022: 594−604
15
马帅, 刘建伟, 左信. 图神经网络综述[J]. 计算机研究与发展, 2022, 59(1): 47−80
Ma Shuai, Liu Jianwei, Zuo Xin. Survey on graph neural network[J]. Journal of Computer Research and Development, 2022, 59(1): 47−80 (in Chinese)
16
Velickovic P, Cucurull G, Casanova A, et al. Graph attention networks[C/OL]//Proc of the 6th Int Conf on Learning Representations. Washington: ICLR. 2018[2025-06-26]. https://openreview.net/forum?id=rJXMpikCZ
17
You Yunning, Chen Tianlong, Sui Yongduo, et al. Graph contrastive learning with augmentations[C]//Proc of the 34th Advances in Neural Information Processing Systems. Cambridge, MA: MIT, 2020, 5812−5823
18
Qiang Qianyao, Zhang Bin, Wang Fei, et al. Fast multi-view discrete clustering with anchor graphs[C]//Proc of the 35th AAAI Conf on Artificial Intelligence. Palo Alto, CA: AAAI, 2021: 9360−9367
19
Kipf T N, Welling M. Variational graph auto-encoders[J]. arXiv preprint, arXiv: 1611.07308, 2016
20
Wang Chun, Pan Shirui, Hu Ruiqi, et al. Attributed graph clustering: A deep attentional embedding approach[C]//Proc of the 28th Int Joint Conf on Artificial Intelligence. San Francisco, CA: Morgan Kaufmann, 2019: 3670−3676
21
Bo Deyu, Wang Xiao, Shi Chuan, et al. Structural deep clustering network[C]//Proc of the 29th Int Conf on World Wide Web. New York: ACM, 2020: 1400−1410
22
Huo Guangyu, Zhang Yong, Gao Junbin, et al. CaEGCN: Cross-attention fusion based enhanced graph convolutional network for clustering[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(4): 3471−3483
23
Li Jintang, Wu Ruofan, Sun Wangbin, et al. What’s behind the mask: Understanding masked graph modeling for graph autoencoders[C]//Proc of the 29th ACM SIGKDD Conf on Knowledge Discovery and Data Mining. New York: ACM, 2023: 1268−1279
24
Bai Lu, Xu Zhuo, Cui Lixin, et al. HC-GAE: The hierarchical cluster-based graph auto-encoder for graph representation learning[C]//Proc of the 38th Advances in Neural Information Proc Systems. Cambridge, MA: MIT, 2024: 127968−127986
25
Yin Hao, Benson A R, Leskovec J, et al. Local higher-order graph clustering[C]//Proc of the 23rd ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2017: 555−564
26
Huang Ling, Wang Changdong, Chao Hongyang. HM-Modularity: A harmonic motif modularity approach for multi-layer network community detection[J]. IEEE Transactions on Knowledge and Data Engineering, 2021, 33(6): 2520−2533
27
Yu Yanlei, Lu Zhiwu, Liu Jiajun, et al. RUM: Network representation learning using motifs[C]//Proc of the 35th Int Conf on Data Engineering. Piscataway, NJ: IEEE, 2019: 1382−1393
28
Dareddy M R, Das M, Yang Hao. motif2vec: Motif aware node representation learning for heterogeneous networks[C]//Proc of the 2019 IEEE Int Conf on Big Data. Piscataway, NJ: IEEE, 2019: 1052−1059
29
Wang Lei, Ren Jing, Xu Bo, et al. MODEL: Motif-based deep feature learning for link prediction[J]. IEEE Transactions on Computational Social Systems, 2020, 7(2): 503−516
30
Shao Ping, Yang Yang, Xu Shengyao, et al. Network embedding via motifs[J]. ACM Transactions on Knowledge Discovery from Data, 2022, 16(3): 44: 1−44: 20
31
Rossi R A, Ahmed N K, Koh E, et al. A structural graph representation learning framework[C]//Proc of the 13th Int Conf on Web Search and Data Mining. New York: ACM, 2020: 483−491
32
Lee J B, Rossi R A, Kong X, et al. Graph convolutional networks with motif-based attention[C]//Proc of the 28th ACM Int Conf on Information and Knowledge Management. New York: ACM, 2019: 499−508
33
Sun Mingchen, Yang Mengduo, Li Yingji, et al. Structural-aware motif-based prompt tuning for graph clustering[J]. Information Sciences, 2023, 649: 119643
34
Wang Xiao, Zhu Meiqi, Bo Deyu, et al. AM-GCN: Adaptive multi-channel graph convolutional networks[C]//Proc of the 26th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining. New York: ACM, 2020: 1243−1253
35
Mccallum A K, Nigam K, Rennie J. et al. Automating the construction of Internet portals with machine learning[J]. Information Retrieval, 2000, 3: 127−163
36
Giles C L, Bollacker K D, Lawrence S. CiteSeer: An automatic citation indexing system[C]//Proc of the 3rd ACM conf on Digital libraries. New York: ACM, 1998: 89−98
37
Pei Hongbin, Wei Bingzhe, Chang K, et al. Geom-GCN: Geometric graph convolutional networks[C/OL]//Proc of the 8th Int Conf on Learning Representations. Washington: ICLR. 2020 [2025-06-26]. https://openreview.net/forum?id=S1e2agrFvS
38
李嘉欣. 基于局部模体的社区发现算法研究[D]. 太原: 山西大学, 2023
Li Jiaxin. Research on community detection algorithm based on local motifs[D]. Taiyuan: Shanxi University, 2023 (in Chinese)
39
Bressan M, Leucci S, Panconesi A. Motivo: Fast motif counting via succinct color coding and adaptive sampling[J]. Proceedings of the VLDB Endowment, 2019, 12(11): 1651−1663
40
Yang Chen, Luo Lailong, Lu Yuliang, et al. SGES: A general and space-efficient framework for graphlet counting in graph streams[C]//Proc of the 33rd ACM Int Conf Information and Knowledge Management. New York: ACM, 2024: 2817−2826
2026年第63卷第2期
181
83
0
引用本文
BibTeX
文章信息
doi: 10.7544/issn1000-1239.202550356
  • 接收时间:2025-05-26
  • 首发时间:2026-11-06
  • 出版时间:2026-02-01
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2025-05-26
  • 修回日期:2025-08-18
  • 录用日期:2025-09-02
基金
作者信息
    1山西大学计算机与信息技术学院 太原 030006
    2计算智能与中文信息处理教育部重点实验室(山西大学) 太原 030006
    3山西大学智能信息处理研究所 太原 030006

通讯作者:

郑文萍()
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/10.7544/issn1000-1239.202550356
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数
科
Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)
属
Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏