Melton Joshua, Krishnan Siddharth
IEEE Trans Pattern Anal Mach Intell. 2023 Sep;45(9):11067-11078. doi: 10.1109/TPAMI.2023.3263079. Epub 2023 Aug 7.
Graph neural networks (GNNs) have become effective learning techniques for many downstream network mining tasks including node and graph classification, link prediction, and network reconstruction. However, most GNN methods have been developed for homogeneous networks with only a single type of node and edge. In this work we present muxGNN, a multiplex graph neural network for heterogeneous graphs. To model heterogeneity, we represent graphs as multiplex networks consisting of a set of relation layer graphs and a coupling graph that links node instantiations across multiple relations. We parameterize relation-specific representations of nodes and design a novel coupling attention mechanism that models the importance of multi-relational contexts for different types of nodes and edges in heterogeneous graphs. We further develop two complementary coupling structures: node invariant coupling suitable for node- and graph-level tasks, and node equivariant coupling suitable for link-level tasks. Extensive experiments conducted on six real-world datasets for link prediction in both transductive and inductive contexts and graph classification demonstrate the superior performance of muxGNN over state-of-the-art heterogeneous GNNs. In addition, we show that muxGNN's coupling attention discovers interpretable connections between different relations in heterogeneous networks.
图神经网络（GNN）已成为许多下游网络挖掘任务的有效学习技术，包括节点和图分类、链接预测以及网络重建。然而，大多数GNN方法是针对仅具有单一类型节点和边的同构网络开发的。在这项工作中，我们提出了muxGNN，一种用于异构图的多重图神经网络。为了对异质性进行建模，我们将图表示为由一组关系层图和一个耦合图组成的多重网络，该耦合图跨多个关系链接节点实例。我们对节点的特定关系表示进行参数化，并设计了一种新颖的耦合注意力机制，该机制对异构图中不同类型的节点和边的多关系上下文的重要性进行建模。我们进一步开发了两种互补的耦合结构：适用于节点级和图级任务的节点不变耦合，以及适用于链接级任务的节点等变耦合。在六个真实世界数据集上进行的广泛实验，用于在转导和归纳上下文中进行链接预测以及图分类，证明了muxGNN优于现有最先进的异质GNN。此外，我们表明muxGNN的耦合注意力发现了异质网络中不同关系之间可解释的连接。
