聚合高阶邻居节点的异构图神经网络模型研究  

Research on Heterogeneous Graph Neural Network Based on Higher-order Neighbors

作　　者：谭鑫媛 裴颂文[1,2] TAN Xin-yuan;PEI Song-wen(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;State Key Lab of Computer Architecture,Chinese Academy of Sciences,Beijing 100190,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]中国科学院计算机体系结构国家重点实验室,北京100190

出　　处：《小型微型计算机系统》2023年第9期1954-1960,共7页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61975124)资助;上海自然科学基金项目(20ZR1438500)资助;上海市科委科技行动计划项目(20DZ2308700)资助;中国计算机体系结构国家重点实验室开放课题项目(CARCHA202111)资助;上海市经信委软件和集成电路产业发展专项项目(RX-RJJC-02-20-4212)资助。

摘　　要：降低异构图的语义和结构信息至低维空间,是解决异构图数据难以高效输入机器学习算法的关键问题.然而,现有的异构图神经网络选择忽视高阶邻居节点,避免学习复杂的结构信息.因此,本文提出一种聚合高阶邻居节点的异构图神经网络模型(HONG).首先提出了基于元路径的高阶邻居子图和面向异构图的池化层HetRepPool,采用GCN学习复杂的结构信息;其次采用HAN学习基于元路径的语义信息;最后通过注意力机制得到节点的嵌入表示,实现了异构图嵌入目的.实验结果表明,HONG与其他图神经网络(GCN、GAT、GraphSAGE、HetGNN、HAN、GAHNE)相比,对于异构图节点分类任务,Micro F1平均提升了3.88%,Macro F1平均提升了4.13%;对于异构图节点聚类任务,ARI平均提升了12.66%,NMI平均提升了12.02%.Embedding the semantic and structural information of heterogeneous graphs into low dimensional space is the key to solve the problem of putting heterogeneous graph data into machine learning algorithm efficiently.However,the existing heterogeneous graph neural networks ignore higher-order neighbors and avoid learning complex structural information.Therefore,this paper proposes a heterogeneous graph neural network model(HONG)that aggregates higher-order neighbors.Firstly,a higher-order subgraph based on meta-path and a pooling operator HetRepPool for heterogeneous graphs are proposed,and combined with GCN to learn complex structural information.Secondly,HAN is used to learn semantic information based on meta-path.Finally,the embedded representation of nodes is obtained through the attention mechanism to achieve heterogeneous graph embedding.The experimental results show that compared with other graph neural networks(GCN、GAT、GraphSAGE、HetGNN、HAN、GAHNE),HONG increase the average Micro F1 by 3.88%and Macro F1 by 4.13%in heterogeneous graph node classification task,the average ARI by 12.66%and NMI by 12.02%in heterogeneous graph node clustering task.

关 键 词：图神经网络 异构图 图嵌入 图池化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]