基于拓扑信息增强的异构图神经网络算法

姚波武; 邓琨; 魏振华; 吴桐; 刘星妍

doi:10.11959/j.issn.1000-0801.2025208

您当前的位置：

首页 >

文章列表页 >

基于拓扑信息增强的异构图神经网络算法

研究与开发 | 更新时间：2026-01-08

- 基于拓扑信息增强的异构图神经网络算法
- Heterogeneous graph neural network algorithm based on topology information enhancement
- 电信科学 2025年41卷第12期页码：128-145
- 作者机构：
  
  1.东华理工大学人工智能与信息工程学院，江西南昌 330013
  2.嘉兴大学浙江省全省多模态感知与智能系统重点实验室，浙江嘉兴 314001
  3.嘉兴大学信息科学与工程学院，浙江嘉兴 314001
  4.嘉兴大学人工智能学院，浙江嘉兴 314001
- 作者简介：
  
  [ "姚波武（2000- ），男，东华理工大学信息工程学院硕士生，主要研究方向为异构图神经网络和深度学习等。" ]
  [ "邓琨（1980- ），男，博士，嘉兴大学浙江省全省多模态感知与智能系统重点实验室、嘉兴大学信息科学与工程学院副教授、硕士生导师，主要研究方向为网络结构分析、数据挖掘、异构网络分析等。" ]
  [ "魏振华（1981- ），女，博士，东华理工大学副教授，主要研究方向为智能信息处理和空间信息管理。" ]
  [ "吴桐（1992- ），男，博士，嘉兴大学浙江省全省多模态感知与智能系统重点实验室、嘉兴大学信息科学与工程学院讲师，主要研究方向为太赫兹/毫米波传感、雷达信号处理、纳米光子学等。" ]
  [ "刘星妍（1980- ），女，嘉兴大学人工智能学院高级工程师，主要研究方向为数据挖掘、网络结构分析等。" ]
- 基金信息：
  
  国家自然科学基金资助项目(62302197);嘉兴市科技公益项目(2024AY10023)
- DOI：10.11959/j.issn.1000-0801.2025208
  中图分类号： TP391
- 收稿：2025-01-24，
  
  修回：2025-05-16，
  
  录用：2025-05-16，
  
  纸质出版：2025-12-20
- 稿件说明：
移动端阅览
姚波武,邓琨,魏振华等.基于拓扑信息增强的异构图神经网络算法[J].电信科学,2025,41(12):128-145.

YAO Bowu,DENG Kun,WEI Zhenhua,et al.Heterogeneous graph neural network algorithm based on topology information enhancement[J].Telecommunications Science,2025,41(12):128-145.
姚波武,邓琨,魏振华等.基于拓扑信息增强的异构图神经网络算法[J].电信科学,2025,41(12):128-145. DOI： 10.11959/j.issn.1000-0801.2025208.

YAO Bowu,DENG Kun,WEI Zhenhua,et al.Heterogeneous graph neural network algorithm based on topology information enhancement[J].Telecommunications Science,2025,41(12):128-145. DOI： 10.11959/j.issn.1000-0801.2025208.

摘要

现有异构图神经网络算法多侧重于节点属性信息，利用注意力机制区分节点的重要性；然而，此类方法在捕获完整拓扑信息和节点特征多样性方面存在不足。针对以上问题，提出一种基于拓扑信息增强的异构图神经网络算法。首先，该算法从局部视角进行聚合，引入局部拓扑信息；随后，结合上下文采样与结构注意力机制，动态识别并加权聚合关键的高阶拓扑信息；最后，通过特征空间叠加技术保留节点的异构信息，并利用多头自注意力机制跨类型聚合节点属性，实现复杂语义的捕获。多个公开数据集上的实验结果表明，该算法在捕获异构图拓扑信息和保留节点异构信息方面表现优异。相较于多种基线方法，在节点分类任务中，Macro-F1指标平均提升0.52%~2.15%；在聚类任务中，归一化互信息（normalized mutual information，NMI）值平均提升1.26%~2.65%。

Abstract

Existing heterogeneous graph neural network algorithms primarily focus on node attribute information

utilizing attention mechanisms to distinguish node importance. However

such methods have limitations in capturing complete topological information and node feature diversity. To address these issues

a heterogeneous graph neural network algorithm based on topological information enhancement was proposed. Firstly

local topological information was incorporated

aggregating from a local perspective. Then

context sampling was combined with structural attention mechanisms to dynamically identify and weight aggregate key higher-order topological information. Finally

node heterogeneous information was preserved through feature space stacking technology and multi-head self-attention mechanisms were employed to cross-type aggregate node attributes

achieving the capture of complex semantics. Experimental results on multiple public datasets demonstrate that this method excels in capturing heterogeneous graph topological information and preserving node heterogeneous information. Compared to various baseline methods

in node classification tasks

the Macro-F1 metric is improved by an average of 0.52%~2.15%. In clustering tasks

the normalized mutual information (NMI) value is improved by an average of 1.26%~2.65%.

关键词

Keywords

references

WANG X , BO D Y , SHI C , et al . A survey on heterogeneous graph embedding: methods, techniques, applications and sources [J ] . IEEE Transactions on Big Data , 2022 , 9 ( 2 ): 415 - 436 .

陶涛 , 李珍 , 王冀彬 , 等 . 基于图神经网络的权益推荐技术方案研究 [J ] . 电信科学 , 2023 , 39 ( 8 ): 91 - 101 .

TAO T , LI Z , WANG J B , et al . Research on the graphical convolution neural network based benefits recommendation system strategy [J ] . Telecommunications Science , 2023 , 39 ( 8 ): 91 - 101 .

LI S B , GONG J , KE S N , et al . Graph transformer-based heterogeneous graph neural networks enhanced by multiple meta-path adjacency matrices decomposition [J ] . Neurocomputing , 2025 , 629 : 129604 .

DONG Y X , CHAWLA N V , SWAMI A . Metapath2vec: scalable representation learning for heterogeneous networks [C ] // Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining . New York : ACM Press , 2017 : 135 - 144 .

JIN D , HUO C Y , LIANG C D , et al . Heterogeneous graph neural network via attribute completion [C ] // Proceedings of the Web Conference 2021 . New York : ACM Press , 2021 : 391 - 400 .

LYU Q S , DING M , LIU Q , et al . Are we really making much progress?: revisiting, benchmarking and refining heterogeneous graph neural networks [C ] // Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining . New York : ACM Press , 2021 : 1150 - 1160 .

LI C , FU J H , YAN Y Y , et al . Higher order heterogeneous graph neural network based on node attribute enhancement [J ] . Expert Systems with Applications , 2024 , 238 : 122404 .

贾霄生 , 赵中英 , 李超 , 等 . 互信息与多条元路径融合的异质网络表示学习方法 [J ] . 软件学报 , 2023 , 34 ( 7 ): 3256 - 3271 .

JIA X S , ZHAO Z Y , LI C , et al . Heterogeneous network representation learning method fusing mutual information and multiple meta-paths [J ] . Journal of Software , 2023 , 34 ( 7 ): 3256 - 3271 .

王瑞琴 , 黄熠旻 , 纪其顺 , 等 . 注意力感知的边-节点交换图神经网络模型 [J ] . 电信科学 , 2024 , 40 ( 1 ): 106 - 114 .

WANG R Q , HUANG Y M , JI Q S , et al . Attention aware edge-node exchange graph neural network [J ] . Telecommunications Science , 2024 , 40 ( 1 ): 106 - 114 .

WANG X , JI H Y , SHI C , et al . Heterogeneous graph attention network [C ] // Proceedings of the World Wide Web Conference . New York : ACM Press , 2019 : 2022 - 2032 .

FU X Y , ZHANG J N , MENG Z Q , et al . MAGNN: metapath aggregated graph neural network for heterogeneous graph embedding [C ] // Proceedings of the Proceedings of The Web Conference 2020 . New York : ACM Press , 2020 : 2331 - 2341 .

LIN W H , CHEN Z L , CHEN Y H , et al . Heterogeneous graph neural network with adaptive relation reconstruction [J ] . Neural Networks , 2025 , 187 : 107313 .

NIKOLENTZOS G , MELADIANOS P , VAZIRGIANNIS M . Matching node embeddings for graph similarity [C ] // Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence . New York : ACM Press , 2017 : 2429 - 2435 .

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 . New York : ACM Press , 2014 : 701 - 710 .

TANG J , QU M , WANG M Z , et al . LINE: large-scale information network embedding [C ] // Proceedings of the 24th International Conference on World Wide Web . New York : ACM Press , 2015 : 1067 - 1077 .

RIBEIRO L F R , SAVERESE P H P , FIGUEIREDO D R . Struc2vec: learning node representations from structural identity [C ] // Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining . New York : ACM Press , 2017 : 385 - 394 .

KIPF T N , WELLING M . Semi-supervised classification with graph convolutional networks [EB ] . 2016 .

SCHLICHTKRULL M , KIPF T N , BLOEM P , et al . Modeling relational data with graph convolutional networks [C ] // Proceedings of the Semantic Web . Cham : Springer International Publishing , 2018 : 593 - 607 .

CHEN K J , LU H , LIU Z , et al . Heterogeneous graph convolutional network with local influence [J ] . Knowledge-Based Systems , 2022 , 236 : 107699 .

杨帅 , 王瑞琴 , 马辉 . 基于多通道的边学习图卷积网络 [J ] . 电信科学 , 2022 , 38 ( 9 ): 95 - 104 .

YANG S , WANG R Q , MA H . Multi-channel based edge-learning graph convolutional network [J ] . Telecommunications Science , 2022 , 38 ( 9 ): 95 - 104 .

VELICKOVIC P , CUCURULL G , CASANOVA A , et al . Graph attention networks [J ] . Stat , 2017 , 1050 ( 20 ): 10 - 48550 .

ZHAO J N , LI C Z , WEN Q L , et al . Gophormer: ego-graph transformer for node classification [EB ] . 2021 .

ZHAO Y F , XU S X , DUAN H . HGNN-BRFE: heterogeneous graph neural network model based on region feature extraction [J ] . Electronics , 2024 , 13 ( 22 ): 4447 .

CHEN Y J , LIU Y . HeGAE-AC: heterogeneous graph auto-encoder for attribute completion [J ] . Knowledge-Based Systems , 2024 , 287 : 111436 .

VASWANI A , SHAZEER N , PARMAR N , et al . Attention is all you need [J ] . Advances in Neural Information Processing Systems , 2017 .

MCPHERSON M , SMITH-LOVIN L , COOK J M . Birds of a feather: homophily in social networks [J ] . Annual Review of Sociology , 2001 , 27 : 415 - 444 .

MAO Q H , LIU Z M , LIU C H , et al . HINormer: representation learning on heterogeneous information networks with graph transformer [C ] // Proceedings of the ACM Web Conference 2023 . New York : ACM Press , 2023 : 599 - 610 .

ZHOU Z A , SHI J M , YANG R C , et al . SlotGAT: slot-based message passing for heterogeneous graph neural network [EB ] . 2024 .

YUN S , JEONG M , KIM R , et al . Graph transformer networks [J ] . Advances in Neural Information Processing Systems , 2019 , 32 .

WANG S H , CAO G T , CAO W M , et al . NLA-GNN: non-local information aggregated graph neural network for heterogeneous graph embedding [J ] . Pattern Recognition , 2025 , 158 : 110940 .

VAN DER MAATEN L , HINTON G . Visualizing data using t-SNE [J ] . Journal of Machine Learning Research , 2008 , 9 ( 11 ).

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于元路径卷积的异构图神经网络算法

注意力感知的边-节点交换图神经网络模型

基于移动通信网信号的人体动作识别研究

基于多通道超图卷积与跨视图对比学习的会话推荐

基于图卷积神经网络的分解可解释性的蜂窝网络流量预测模型