计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (3): 188-196.doi: 10.11896/jsjkx.250600067

• 数据库 & 大数据 & 数据科学 • 上一篇    下一篇

基于KAN的双通道图神经网络

王静红1,2,3,4, 李鹏超1,3,4,5, 王熙照6, 张自立1,3,4,5   

  1. 1 河北师范大学计算机与网络空间安全学院 石家庄 050024
    2 河北工程技术学院人工智能学院 石家庄 050020
    3 中国科学技术大学认知智能全国重点实验室 合肥 230088
    4 河北省网络与信息安全重点实验室 石家庄 050024
    5 供应链大数据分析与数据安全河北省工程研究中心 石家庄 050024
    6 深圳大学计算机与软件学院 广东 深圳 518060
  • 收稿日期:2025-06-11 修回日期:2025-08-21 发布日期:2026-03-12
  • 通讯作者: 王静红(wangjinghong@126.com)
  • 基金资助:
    河北省自然科学基金(F2024205028);河北省研究生创新资助项目(CXZZSS2025049);河北师范大学科技类基金(L2023J05);认知智能全国重点实验室开放课题(COGOS-2025HE07)

Dual-channel Graph Neural Network Based on KAN

WANG Jinghong1,2,3,4, LI Pengchao1,3,4,5, WANG Xizhao6, ZHANG Zili1,3,4,5   

  1. 1 College of Computer and Cyber Security, Hebei Normal University, Shijiazhuang 050024, China
    2 College of Artificial Intelligence, Hebei University of Engineering Science, Shijiazhuang 050020, China
    3 State Key Laboratory of Cognitive Intelligence, University of Science and Technology of China, Hefei 230088, China
    4 Hebei Provincial Key Laboratory of Network and Information Security, Shijiazhuang 050024, China
    5 Hebei Provincial Engineering Research Center for Supply Chain Big Data Analytics & Data Security, Shijiazhuang 050024, China
    6 College of Computer Science and Software Engineering, Shenzhen University, Shenzhen, Guangdong 518060, China
  • Received:2025-06-11 Revised:2025-08-21 Online:2026-03-12
  • About author:WANG Jinghong,born in 1967,Ph.D,professor,is a member of CCF(No.58341S).Her main research interests include artificial intelligence,pattern recognition,machine learning and data mining.
  • Supported by:
    Natural Science Foundation of Hebei Province(F2024205028),Postgraduate’s Innovation Fund Project of Hebei Province(CXZZSS2025049),Hebei Normal University Science and Technology Fund(L2023J05) and Open Fund of State Key Laboratory of Cognitive Intelligence,iFLYTEK(COGOS-2025HE07).

PDF (PC)

122

摘要： 图神经网络(GNNs)是一种专门针对图数据的神经网络模型,近年来被成功应用在各种图学习任务上,如节点分类、链路预测等。然而,目前的图神经网络模型大多基于消息传递范式,无法充分捕捉节点的结构信息与特征信息之间的多维关联关系。此外,传统激活函数容易导致信息丢失和模型解释性不足的问题。为此,提出了一种基于Kolmogorov-Arnold网络(KAN)的双通道图神经网络(KDCGNN)。KDCGNN利用结构卷积和特征卷积,从两个通道分别提取图的结构信息和特征信息,生成节点的结构编码和特征编码,拼接融合后,进一步借助KAN对嵌入表示进行特征转换,提升分类性能和模型的可解释性。同时,引入一致性损失函数,鼓励结构编码和特征编码之间的分布一致性,从而增强模型的泛化能力。在3个经典引文网络数据集(Cora,Citeseer,Pubmed)上的实验表明,KDCGNN在节点分类任务中的表现优于现有基准方法。KDCGNN的提出为图神经网络的可解释性与性能优化提供了新思路。

关键词: 图神经网络, Kolmogorov-Arnold网络, 双通道机制, 节点分类, 高斯-Dice相似度

Abstract: GNNs are specialized models designed for graph data and have been successfully applied to various graph learning tasks such as node classification and link prediction.However,most existing GNN models are based on the message-passing paradigm,which fails to fully capture the multi-dimensional relationships between structural information and feature information of nodes.Additionally,traditional activation functions often lead to information loss and lack interpretability in the models.To address these challenges,this paper proposes a novel Kolmogorov-Arnold Network-based Dual-Channel Graph Neural Network(KDCGNN).KDCGNN employs structural convolution and feature convolution in two separate channels to extract structural and feature information from graphs,generating structural and feature encodings for nodes.These encodings are then fused through concatenation and further transformed using the Kolmogorov-Arnold Network to enhance classification performance and model interpretability.Furthermore,a consistency loss function is introduced to encourage distributional alignment between structural and feature encodings,thereby improving the generalization capability of the model.Experiments on three benchmark citation network datasets(Cora,Citeseer,and Pubmed) demonstrate that KDCGNN outperforms existing baseline methods in node classification tasks.KDCGNN provides a novel approach to improving the interpretability and performance of graph neural networks.

Key words: Graph neural networks, Kolmogorov-Arnold networks, Dual-channel mechanism, Node classification, Gaussian-Dice similarity

中图分类号: 

  • TP391
[1]NIU M,WANG C,ZHANG Z,et al.A computational model of circRNA-associated diseases based on a graph neural network:prediction and case studies for follow-up experimental validation[J].BMC Biology,2024,22(1):24.
[2]NIU M,ZOU Q,LIN C.CRBPDL:Identification of circRNA-RBP interaction sites using an ensemble neural network approach[J].PLoS Computational Biology,2022,18(1):e1009798.
[3]KIPF T N,WELLING M.Semi-supervised classification withgraph convolutional networks[J].arXiv:1609.02907,2016.
[4]VELICKOVIC P,CUCURULL G,CASANOVA A,et al.Graph attention networks[J].arXiv:1710.10903,2017.
[5]LIU Z,WANG Y,VAIDYA S,et al.Kan:Kolmogorov-arnold networks[J].arXiv:2404.19756,2024.
[6]CYBENKO G.Approximation by superpositions of a sigmoidal function[J].Mathematics of Control,Signals and Systems,1989,2(4):303-314.
[7]HORNIK K,STINCHCOMBE M,WHITE H.Multilayer feedforward networks are universal approximators[J].Neural Networks,1989,2(5):359-366.
[8]ZHANG F,ZHANG X.GraphKAN:Enhancing Feature Extraction with Graph Kolmogorov Arnold Networks[J].arXiv:2406.13597,2024.
[9]KIAMARI M,KRISHNAMACHARI B.GKAN:Graph Kol-mogorov-Arnold Networks[J].arXiv:2406.06470,2024.
[10]BRESSON R,NIKOLENTZOS G,PANAGOPOULOS G,et al.Kagnns:Kolmogorov-arnold networks meet graph learning[J].arXiv:2406.18380,2024.
[11]DE CARLO G,MASTROPIETRO A,ANAGNOSTOPOULOS A.Kolmogorov-arnold graph neural networks[J].arXiv:2406.18354,2024.
[12]AHMED T,SIFAT M H R.Graphkan:Graph kolmogorov arnold network for small molecule-protein interaction predictions[C]//ICML’24 Workshop ML for Life and Material Science:From Theory to Industry Applications.2024.
[13]ZHOU K,SONG Q,HUANG X,et al.Multi-channel graph convolutional networks[J].arXiv:1912.08306,2019.
[14]ZHAO R,SHAO Z,ZHANG W,et al.A multi-channel multi-tower GNN model for job transfer prediction based on academic social network[J].Applied Soft Computing,2023,142:110300.
[15]MENG L,YE Z,YANG Y,et al.DeepMCGCN:Multi-channel Deep Graph Neural Networks[J].International Journal of Computational Intelligence Systems,2024,17(1):41.
[16]WANG R,LI F,LIU S,et al.Adaptive Multi-Channel DeepGraph Neural Networks[J].Symmetry,2024,16(4):406.
[17]SEN P,NAMATA G,BILGIC M,et al.Collective classification in network data[J].AI Magazine,2008,29(3):93-93.
[18]GROVER A,LESKOVEC J.node2vec:Scalable feature learning for networks[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mi-ning.2016:855-864.
[19]PEROZZI B,AL-RFOU R,SKIENA S.Deepwalk:Online learning of social representations[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2014:701-710.
[20]HAMLTON W,YING Z,LESKOVEC J.Inductive representa-tion learning on large graphs[C]//Advances in Neural Information Processing Systems.2017.
[21]WU F,SOUZA A,ZHANG T,et al.Simplifying graph convolutional networks[C]//International Conference on Machine Learning.PMLR,2019:6861-6871.
[22]ZHU J,ROSSI R A,RAO A,et al.Graph neural networks with heterophily[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2021:11168-11176.
[23]WANG T,JIN D,WANG R,et al.Powerful graph convolutional networks with adaptive propagation mechanism for homophily and heterophily[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2022:4210-4218.
[24]GAO H,JI S.Graph U-Nets[C]//International Conference on Machine Learning.PMLR,2019:2083-2092.
[25]WU J,CHEN X,XU K,et al.Structural entropy guided graph hierarchical pooling[C]//International Conference on Machine Learning.PMLR,2022:24017-24030.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市两江新区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发