计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (7): 110-118.doi: 10.11896/jsjkx.240400093

• 计算机软件 • 上一篇    下一篇

融合位置和结构信息的图神经网络的节点学习研究

郝佳辉, 万源, 张宇航   

  1. 武汉理工大学理学院 武汉 430070
  • 收稿日期:2024-04-15 修回日期:2024-09-05 发布日期:2025-07-17
  • 通讯作者: 万源(wanyuan@whut.edu.cn)
  • 作者简介:(281797@whut.edu.cn)
  • 基金资助:
    中央高校基本科研业务费专项资金(2021III030JC)

Research on Node Learning of Graph Neural Networks Fusing Positional and StructuralInformation

HAO Jiahui, WAN Yuan, ZHANG Yuhang   

  1. School of Science, Wuhan University of Technology, Wuhan 430070, China
  • Received:2024-04-15 Revised:2024-09-05 Published:2025-07-17
  • About author:HAO Jiahui,born in 1999,postgra-duate.His main research interests include data mining,machine learning and pattern recognition.
    WAN Yuan,born in 1976,Ph.D,professor.Her main research interests include data mining,pattern recognition,manifold learning,machine learning and feature selection.
  • Supported by:
    Fundamental Research Funds for the Central Universities of Ministry of Education of China(2021III030JC).

PDF (PC)

23

摘要： 图神经网络是一种强大的学习图数据的模型,通过节点信息嵌入和图卷积运算实现图结构数据的表示。图数据中节点的结构信息和节点的位置信息对获取图特征至关重要,但现有的图神经网络同时捕获位置信息和结构信息的表达能力有限。对此,提出了一种新的图神经网络——融合位置和结构信息的图神经网络(Positional and Structural Information with Graph Neural Networks,PSI-GNN)。PSI-GNN的核心思想在于利用编码器获取节点的位置和结构信息,并将这些信息特征嵌入到网络中。通过在网络中更新和传递这两种信息,PSI-GNN实现了对位置和结构信息的有效融合与利用,为解决上述问题提供了有效的解决方案。同时,为应对不同类型的图学习任务,PSI-GNN给予位置和结构信息以不同的权重来应对不同的下游任务。为了验证PSI-GNN的有效性,在多个基准图数据集上进行了实验。实验结果表明,PSI-GNN在节点级任务上最高提升了约14%,在图级任务上最高提升了约35%,验证了PSI-GNN在同时捕获位置和结构信息方面的有效性。

关键词: 图神经网络, 位置信息, 结构信息, 拉普拉斯位置编码, Adamic-Adar结构编码

Abstract: Graph neural networks are powerful models for learning graph-structured data,representing them through node information embedding and graph convolution operations.In graph data,the structural information and positional information of nodes are crucial for extracting graph features.However,existing GNNs have limited expressive ability in simultaneously capturing positional and structural information.This paper proposes a novel graph neural network,named Positional and Structural Information with Graph Neural Networks(PSI-GNN).The core idea of PSI-GNN lies in utilizing an encoder to capture the positional and structural information of nodes and embedding these information features into the network.By updating and propagating these two types of information within the network,PSI-GNN effectively integrates and utilizes positional and structural information,providing an effective solution to the aforementioned problem.Additionally,to accommodate different types of graph learning tasks,PSI-GNN assigns varying weights to positional and structural information based on the specific downstream tasks.To validate the effectiveness of PSI-GNN,experiments are conducted on multiple benchmark graph datasets.The experimental results demonstrate that PSI-GNN achieves a maximum improvement of approximately 14% on node-level tasks and approximately 35% on graph-level tasks.These results confirm the effectiveness of PSI-GNN in simultaneously capturing positional and structural information.

Key words: Graph neural networks, Positional information, Structural information, Laplace position encoding, Adamic-Adar structural encoding

中图分类号: 

  • TP391
[1]ZHANG P,CHEN J,CHE C,et al.IEA-GNN:Anchor-aware Graph Neural Network Fused with Information Entropy for Node Classification and Link Prediction[J].Information Sciences,2023,634:665-676.
[2]ZOU M,GAN Z,CAO R,et al.Similarity-navigated Graph Neural Networks for Node Classification[J].Information Sciences,2023,633:41-69.
[3]DENG X,SUN J,LU J.Graph Neural Network-based Efficient Subgraph Embedding Method for Link Prediction in Mobile Edge Computing[J].Sensors,2023,23:4936.
[4]WU F,JING X Y,WEI P,et al.Semi-supervised Multi-viewGraph Convolutional Networks with Application to Webpage Classification[J].Information Sciences,2022,591:142-154.
[5]BRUNA J,ZAREMBA W,SZLAM A,et al.Spectral Networks and Locally Connected Networks on Graphs[C]//International Conference on Learning Representations.Banff,Canada:OpenReview.net,2014:1-14.
[6]GILMER J,SCHOENHOLZ S S,RILEY P F,et al.Neural Message Passing for Quantum Chemistry[C]//International Confe-rence on Machine Learning.Sydney,Australia:ACM,2017:1263-1272.
[7]XU K,HU W,LESKOVEC J,et al.How Powerful are GraphNeural Networks?[C]//International Conference on Learning Representations.Vancouver,Canada:OpenReview.net,2018:1-17.
[8]MURPHY R L,SRINIVASAN B,RAO V,et al.Relational Pooling for Graph Representations[C]//International Conference on Machine Learning.Los Angeles,USA:ACM,2019:4663-4673.
[9]YOU J X,YING R X,LESKOVEC J.Position-aware GraphNeural Networks[C]//International Conference on Machine Learning.Los Angeles,USA:ACM,2019:7134-7143.
[10]YOU J X,GOMES-SELMAN J,YING R,et al.Identity-aware Graph Neural Networks[C]//AAAI Conference on Artificial Intelligence.Vancouver,Canada:AAAI,2021:10737-10745.
[11]CHEN Y,YOU J,HE J,et al.SP-GNN:Learning Structure and Position Information from Graphs[J].Neural Networks,2023,161:505-514.
[12]VELICKOVIC P,CUCURULL G,CASANOVA A,et al.Graph Attention Networks[C]//International Conference on Learning Representations.Vancouver,Canada:OpenReview.net,2018:1-12.
[13]HAMILTON W L,YING Z,LESKOVEC J.Inductive Repre-sentation Learning on Large Graphs[C]//Annual Conference on Neural Information Processing Systems.Long Beach,USA:MIT Press,2017:1025-1035.
[14]CHEN M,WEI Z,HUANG Z,et al.Simple and Deep Graph Convolutional Networks[C]//International Conference on Machine Learning.Vienna,Austria:ACM,2020.
[15]FAN X,GONG M,WU Y,et al.Propagation Enhanced Neural Message Passing for Graph Representation Learning[J].IEEE Transactions on Knowledge and Data Engineering,2021,35(2):1952-1964.
[16]VASWANI A,SHAZEER N,PARMAR N,et al.Attention is All you Need[C]//Annual Conference on Neural Information Processing Systems.Long Beach,USA:MIT Press,2017:5998-6008.
[17]LI P,WANG Y,WANG H,et al.Distance Encoding:Design Provably More Powerful Neural Networks for Graph Representation Learning[C]//Annual Conference on Neural Information Processing Systems.Vancouver,Canada:MIT Press,2020:4465-4478.
[18]NISHAD S,AGARWAL S,BHATTACHARYA A,et al.GraphReach:Position-aware Graph Neural Network using Reachability Estimations[C]//International Joint Conference on Artificial Intelligence.Montréal,Canada:Morgan Kaufmann,2021:1527-1533.
[19]DWIVEDI V P,JOSHI C K,LAURENT T,et al.Benchmarking Graph Neural Networks[J].Journal of Machine Learning Research,2023,24(43):1-48.
[20]WANG D,CUI P,ZHU W.Structural Deep Network Embedding[C]//ACM SIGKDD International Conference on Know-ledge Discovery and Data Mining.San Francisco,USA:ACM,2016:1225-1234.
[21]RIBEIRO L F R,SAVERESE P H P,FIGUEIREDO D R.struc2vec:Learning Node Representations from Structural Identity[C]//ACM SIGKDD International Conference on Know-ledge Discovery and Data Mining.Halifax,Canada:ACM,2017:385-394.
[22]BAI W,GUO C.Recommendation Based on Graph Neural Network with Structural Identity[C]//International Conference on Machine Learning.Baltimore,USA:ACM,2022:127-131
[23]DWIVEDI V,LUU A,LAURENT T,et al.Graph Neural Networks with Learnable Structural and Positional Representations[C]//International Conference on Learning Representations.Virtual:OpenReview.net,2022:1-25.
[24]YAN Y,HASHEMI M,SWERSKY K,et al.Two Sides of the Same Coin:Heterophily and Oversmoothing in Graph Convolutional Neural Networks[C]//International Conference on Data Mining.Orlando,USA:IEEE,2022:1287-1292.
[25]OU M,CUI P,PEI J,et al.Asymmetric Transitivity Preserving Graph Embedding[C]//ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.San Francisco,USA:ACM,2016:1225-1234.
[26]SEN P,NAMATA G,BILGIC M,et al.Collective Classification in Network Data[J].AI Magazine,2008,29(3):93-106.
[27]HU W,FEY M,ZITNIK M,et al.Open Graph Benchmark:Datasets for Machine Learning on Graphs[C]//Annual Conference on Neural Information Processing Systems.Vancouver,Canada:MIT Press,2020:4465-4478.
[28]IRWIN J J,STERLING T,MYSINGER M M,et al.ZINC:A Free Tool to Discover Chemistry for Biology[J].Journal of Chemical Information and Modeling,2012,52(7):1757-1768.
[29]ROZEMBERCZKI B,KISS O,SARKAR R.Karate Club:AnAPI Oriented Open-source Python Framework for Unsupervised Learning on Graphs[C]//International Conference on Information and Knowledge Management.Los Angeles,Electr Network,2020:3125-3132.
[30]BRESSON X,LAURENT T.Residual Gated Graph ConvNets[J].arXiv:1711.07553,2018.
[31]KIPF T,WELLING M.Semi-Supervised Classification withGraph Convolutional Networks[C]//International Conference on Learning Representations.Toulon,France:OpenReview.net,2017:1-14.
[32]DWIVEDI V,BRESSON X.A Generalization of TransformerNetworks to Graphs[J].arXiv:2012.09699v2,2021.
[33]KREUZER D,BEAINI D,HAMILTON W L,et al.Rethinking Graph Transformers with Spectral Attention[C]//Annual Conference on Neural Information Processing Systems.Sydney,Australia:MIT Press,2021:1-12.
[34]YING C,CAI T,LUO S J,et al.Do Transformers Really Perform Badly for Graph Representation[C]//Annual Conference on Neural Information Processing Systems.Sydney,Australia:MIT Press,2021:1-19.
[35]LI G,XIONG C,THABET A,et al.DeeperGCN:All You Need to Train Deeper GCNs[J].arXiv:2006.07739,2022.
[36]CORSO G,CAVALLERI L,BEAINI D,et al.Principal Neighbourhood Aggregation for Graph Nets[C]//Annual Conference on Neural Information Processing Systems.Vancouver,Canada:MIT Press,2020:13260-13271.
[37]BEAINI D,PASSARO S,LÉTOURNEAU V,et al.Directional Graph Networks[C]//International Conference on Machine Learning.PMLR,2021:748-758.
[38]TIAN Y,ZHANG C,GUO Z,et al.Learning MLPs on Graphs:A Unified View of Effectiveness,Robustness,and Efficiency[C]//International Conference on Learning Representations.Kigali,Africa:OpenReview.net,2023:1-12.
[39]MARTINKUS K,PAPP P,SCHESCH B,et al.Agent-basedGraph Neural Networks[C]//International Conference on Learning Representations.Kigali,Africa:OpenReview.net,2023:1-32.
[40]LIM D,ROBINSON J,ZHAO L,et al.Sign and Basis Invariant Networks for Spectral Graph Representation Learning[C]//International Conference on Learning Representations.Kigali,Africa:OpenReview.net,2023:1-42.
[41]GRAVINA A,BACCIU D,GALLICCHIO C.Anti-SymmetricDGN:A Stable Architecture for Deep Graph Networks[C]//International Conference on Learning Representations.Kigali,Africa:OpenReview.net,2023:1-19.
Viewed
Full text


Abstract

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