Abstract:Graph convolutional network (GCN) nowadays become new state-of-the-art for networks representation learning. Most of the existing methods are single-granular methods that failed to analyze the graph at multi-granular views so as to lose abundant information. Advanced graph pooling techniques can be successfully benefiting from semi-supervised networks representation learning. How to capture multi-granular information through the graph pooling techniques on graphs without additional input features is a great challenge. Technically speaking, we propose our graph node embeddings framework, MGPOOL. First, inspired by the triadic influence learning, we use the 3-clique algorithm to coarsen the graph repeatedly. Three nodes of a triangle form a supernode. We treat the supernodes as key nodes for our graph pooling operations. That keeps the local relationship. These graphs capture consecutive 3-cliques from the finest to the coarsest to preserve global structural relationships. Second, we use the unsupervised single-granular algorithms on the coarsest graph to acquire its node embeddings. Based on that, our graph pooling operations combining with that node embeddings to generate another same size of the coarsest graph. This makes up for the uniqueness of the coarsening result at a time and expands the receptive field for each node to avoid high-proximity information lost. Third, we take the embeddings, the coarsest graph and new coarsest graph as uniform input of MGPOOL. We restore the coarsest graph to the original graph to get the original graph node embeddings. The experimental results on four public datasets, Wiki, Cora, CiteSeer, and DBLP, demonstrate that our method has a better Macro F1 value for node classification tasks and AUC and Ap value for link prediction than the baseline methods.

Pooling in Graph Convolutional Neural Networks

MGPOOL: multi-granular graph pooling convolutional networks representation learning

Uniform Pooling for Graph Networks

Maximal Independent Sets for Pooling in Graph Neural Networks

SPGNN: Recognizing Salient Subgraph Patterns via Enhanced Graph Convolution and Pooling

GSAPool: Gated Structure Aware Pooling for Graph Representation Learning.

Hierarchical Graph Pooling with Self-Adaptive Cluster Aggregation

Graphon Pooling in Graph Neural Networks

Graph explicit pooling for graph-level representation learning

DiffGCN: Graph Convolutional Networks via Differential Operators and Algebraic Multigrid Pooling

Careful Selection and Thoughtful Discarding: Graph Explicit Pooling Utilizing Discarded Nodes

Haar Graph Pooling

AttPool: Towards Hierarchical Feature Representation in Graph Convolutional Networks Via Attention Mechanism

Feature pyramid-based graph convolutional neural network for graph classification

ABDPool: Attention-based Differentiable Pooling.

Training Large-Scale Graph Neural Networks Via Graph Partial Pooling

Graph Pooling for Graph Neural Networks: Progress, Challenges, and Opportunities

Graph Analysis and Graph Pooling in the Spatial Domain

Hierarchical Graph Representation Learning with Differentiable Pooling

SSHPool: The Separated Subgraph-based Hierarchical Pooling

Graph Pooling by Local Cluster Selection