sis 23(6):

Research Article

MIED : An Improved Graph Neural Network for Node Embedding in Heterogeneous Graphs

Download235 downloads
  • @ARTICLE{10.4108/eetsis.3824,
        author={Mingjian  Ni and Yinghao Song and Gongju Wang and Lanxiao Feng and Yang Li and Long Yan and Dazhong Li and Yanfei Wang and Shikun Zhang and Yulun Song},
        title={MIED : An Improved Graph Neural Network for Node Embedding in Heterogeneous Graphs},
        journal={EAI Endorsed Transactions on Scalable Information Systems},
        volume={10},
        number={6},
        publisher={EAI},
        journal_a={SIS},
        year={2023},
        month={9},
        keywords={Heterogeneous Graph, Node Embedding, Metapath, Graph Convolutional Network, Exponential Decay Encoder},
        doi={10.4108/eetsis.3824}
    }
    
  • Mingjian Ni
    Yinghao Song
    Gongju Wang
    Lanxiao Feng
    Yang Li
    Long Yan
    Dazhong Li
    Yanfei Wang
    Shikun Zhang
    Yulun Song
    Year: 2023
    MIED : An Improved Graph Neural Network for Node Embedding in Heterogeneous Graphs
    SIS
    EAI
    DOI: 10.4108/eetsis.3824
Mingjian Ni1, Yinghao Song2, Gongju Wang2, Lanxiao Feng3, Yang Li2, Long Yan2, Dazhong Li2, Yanfei Wang4, Shikun Zhang1, Yulun Song3,*
  • 1: Peking University
  • 2: China Unicom Digital Technology Company Limited
  • 3: China Unicom Network Communications Company Limited
  • 4: Unicom (Beijing) Industrial Internet Company Limited
*Contact email: songyl100@chinaunicom.cn

Abstract

This paper proposes a Metapath-Infused Exponential Decay graph neural network (MIED) approach for node embedding in heterogeneous graphs. It is designed to address limitations in existing methods, which usually lose the graph information during feature alignment and ignore the different importance of nodes during metapath aggregation. Firstly, graph convolutional network (GCN) is applied on the subgraphs, which is derived from the original graph with given metapaths to transform node features. Secondly, an exponential decay encoder (EDE) is designed, in which the influence of nodes on starting point decays exponentially with a fixed parameter as they move farther away from it. Thirdly, a set of experiments is conducted on two selected datasets of heterogeneous graphs, i.e., IMDb and DBLP, for comparison purposes. The results show that MIED outperforms selected approaches, e.g., GAT, HAN, MAGNN, etc. Thus, our approach is proven to be able to take full advantage of graph information considering node weights based on distance aspects. Finally, relevant parameters are analyzed and the recommended hyperparameter setting is given.