Abstract:While Graph Neural Network (GNN) has shown superiority in learning node representations of homogeneous graphs, leveraging GNN on heterogeneous graphs remains a challenging problem. The dominating reason is that GNN learns node representations by aggregating neighbors' information regardless of node types. Some work is proposed to alleviate such issue by exploiting relations or meta-path to sample neighbors with distinct categories, then use attention mechanism to learn different importance for different categories. However, one limitation is that the learned representations for different types of nodes should own different feature spaces, while all the above work still project node representations into one feature space. Moreover, after exploring massive heterogeneous graphs, we identify a fact, that multiple nodes with the same type always connect to a node with another type, which reveals the many-to-one schema, a.k.a. the hierarchical tree structure. But all the above work cannot preserve such tree structure, since the exact multi-hop path correlation from neighbors to the target node would be erased through aggregation. Therefore, to overcome the limitations of the literature, we propose T-GNN, a tree structure-aware graph neural network model for graph representation learning. Specifically. the proposed T-GNN consists of two modules: (1) the integrated hierarchical aggregation module and (2) the relational metric learning module. The integrated hierarchical aggregation module aims to preserve the tree structure by combining GNN with gated recurrent unit to integrate the hierarchical and sequential neighborhood information on the tree structure to node representations. The relational metric learning module aims to preserve the heterogeneity by embedding each type of nodes into a type-specific space with distinct distribution based on similarity metrics. In this way, our proposed T-GNN is capable of simultaneously preserving the heterogeneity and the tree structure inherent in heterogeneous graphs. Finally, we conduct extensive experiments to show the outstanding performance of T-GNN in tasks of node clustering and classification, inductive node clustering and classification, and link prediction.

Bipartite Graph Neural Networks for Efficient Node Representation Learning

Cascade-BGNN: Toward Efficient Self-supervised Representation Learning on Large-scale Bipartite Graphs

Discovering the Representation Bottleneck of Graph Neural Networks from Multi-order Interactions

Exploiting Node-Feature Bipartite Graph in Graph Convolutional Networks.

Bilinear Graph Neural Network with Neighbor Interactions

Scalable Graph Neural Networks Via Bidirectional Propagation.

Graph Pointer Neural Networks

Geodesic Graph Neural Network for Efficient Graph Representation Learning

ByteGNN: Efficient Graph Neural Network Training at Large Scale

Hierarchical Bipartite Graph Convolution Networks

Hierarchical Message-Passing Graph Neural Networks

Adversarial Representation Learning on Large-Scale Bipartite Graphs

DGNN: Decoupled Graph Neural Networks with Structural Consistency between Attribute and Graph Embedding Representations

Learning Vertex Representations for Bipartite Networks

HighwayGraph: Modelling Long-distance Node Relations for Improving General Graph Neural Network.

DPGNN: Dual-perception graph neural network for representation learning

Tree Structure-Aware Graph Representation Learning Via Integrated Hierarchical Aggregation and Relational Metric Learning

Adversarial Representation Learning on Large-Scale Bipartite Graphs.

Scalable and Efficient Full-Graph GNN Training for Large Graphs

Universal Inceptive GNNs by Eliminating the Smoothness-generalization Dilemma

Bipartite Graph Convolutional Networks with Adversarial Domain Transfer