异质图的属性多阶图卷积网络（英文版）

英文标题：Attributed Multi-order Graph Convolutional Network for Heterogeneous Graphs中文摘要：本篇论文主要介绍了一种基于 AMOGCN 的异构图神经网络，该方法可以学习具有区分度的节点嵌入和关系，并能够自动地构建元路径来影响嵌入的质量，从而提高半监督分类性能。英文摘要：Heterogeneous graph neural networks aim to discover discriminative nodeembeddings and relations from multi-relational networks.One challenge ofheterogeneous graph learning is the design of learnable meta-paths, whichsignificantly influences the quality of learned embeddings.Thus, in this paper,we propose an Attributed Multi-