Cross-Graph Interaction Networks

被引:1
作者
Guo, Qihang [1 ]
Yang, Xibei [1 ]
Ding, Weiping [2 ]
Qian, Yuhua [3 ]
机构
[1] Jiangsu Univ Sci & Technol, Sch Econ & Management, Zhenjiang 212100, Peoples R China
[2] Nantong Univ, Sch Informat Sci & Technol, Nantong 226019, Peoples R China
[3] Shanxi Univ, Inst Big Data Sci & Ind, Taiyuan 030006, Peoples R China
关键词
Topology; Message passing; Graph neural networks; Unsupervised learning; Training; Computer architecture; Robustness; Network topology; Electronic mail; Contrastive learning; cross-graph interaction; inter-graph message passing;
D O I
10.1109/TKDE.2025.3543377
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Graph neural networks (GNNs) are recognized as a significant methodology for handling graph-structure data. However, with the increasing prevalence of learning scenarios involving multiple graphs, traditional GNNs mostly overlook the relationships between nodes across different graphs, mainly due to their limitation of traditional message passing within each graph. In this paper, we propose a novel GNN architecture called cross-graph interaction networks (GInterNet) to enable inter-graph message passing. Specifically, we develop a cross-graph topology construction module to uncover and learn the potential topologies between nodes across different graphs. Furthermore, we establish inter-graph message passing based on the learned cross-graph topologies, achieving cross-graph interaction by aggregating information from different graphs. Finally, we employ cross-graph construction functions involving the relationships between contextual information and cross-graph topology structure to iteratively update the cross-graph topologies. Different to existing related approaches, GInterNet is designed as a cross-graph interaction paradigm for inter-graph message passing. It enables multi-graph interaction during the message passing process. Additionally, it is a plug-and-play framework that can be easily embedded into other models. We evaluate its performance in semi-supervised and unsupervised learning scenarios involving multiple graphs. A detailed theoretical analysis and extensive experiment results have shown that GInterNet improves the performance and robustness of the base models.
引用
收藏
页码:2341 / 2355
页数:15
相关论文
共 50 条
[31]   A Flexible Diffusion Convolution for Graph Neural Networks [J].
Zhao, Songwei ;
Yu, Bo ;
Yang, Kang ;
Zhang, Sinuo ;
Hu, Jifeng ;
Jiang, Yuan ;
Yu, Philip S. ;
Chen, Hechang .
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2025, 37 (06) :3118-3131
[32]   Graph Neural Networks for Wireless Networks: Graph Representation, Architecture and Evaluation [J].
Lu, Yang ;
Li, Yuhang ;
Zhang, Ruichen ;
Chen, Wei ;
Ai, Bo ;
Niyato, Dusit .
IEEE WIRELESS COMMUNICATIONS, 2025, 32 (01) :150-156
[33]   Modular Meta-Learning for Power Control via Random Edge Graph Neural Networks [J].
Nikoloska, Ivana ;
Simeone, Osvaldo .
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, 2023, 22 (01) :457-470
[34]   Collaborative Graph Neural Networks for Attributed Network Embedding [J].
Tan, Qiaoyu ;
Zhang, Xin ;
Huang, Xiao ;
Chen, Hao ;
Li, Jundong ;
Hu, Xia .
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2024, 36 (03) :972-986
[35]   Learning Stochastic Graph Neural Networks With Constrained Variance [J].
Gao, Zhan ;
Isufi, Elvin .
IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2023, 71 :358-371
[36]   Plausible Heterogeneous Graph $k$-Anonymization for Social Networks [J].
Li, Kaiyang ;
Tian, Ling ;
Zheng, Xu ;
Hui, Bei .
TSINGHUA SCIENCE AND TECHNOLOGY, 2022, 27 (06) :912-924
[37]   Multi-Relation Augmentation for Graph Neural Networks [J].
Xiao, Shunxin ;
Lin, Huibin ;
Wang, Jianwen ;
Qin, Xiaolong ;
Wang, Shiping .
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE, 2024, 8 (05) :3614-3627
[38]   Power Flow Balancing With Decentralized Graph Neural Networks [J].
Hansen, Jonas Berg ;
Anfinsen, Stian Normann ;
Bianchi, Filippo Maria .
IEEE TRANSACTIONS ON POWER SYSTEMS, 2023, 38 (03) :2423-2433
[39]   On Inductive-Transductive Learning With Graph Neural Networks [J].
Ciano, Giorgio ;
Rossi, Alberto ;
Bianchini, Monica ;
Scarselli, Franco .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2022, 44 (02) :758-769
[40]   Adaptive Backdoor Attacks With Reasonable Constraints on Graph Neural Networks [J].
Dong, Xuewen ;
Li, Jiachen ;
Li, Shujun ;
You, Zhichao ;
Qu, Qiang ;
Kholodov, Yaroslav ;
Shen, Yulong .
IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, 2025, 22 (04) :4053-4069