Multi-grained Semantics-aware Graph Neural Networks

TL;DR

This paper introduces AdamGNN, a unified graph neural network model that jointly learns node and graph representations by leveraging multi-grained semantics, improving performance on various tasks through innovative pooling and unpooling operators.

Contribution

The work presents a novel AdamGNN model that integrates multi-grained semantics into node and graph representation learning, outperforming existing methods on multiple datasets.

Findings

Significantly outperforms 17 competing models on 14 datasets

Effectively captures long-range interactions in graphs

Ablation studies validate the components' effectiveness

Abstract

Graph Neural Networks (GNNs) are powerful techniques in representation learning for graphs and have been increasingly deployed in a multitude of different applications that involve node- and graph-wise tasks. Most existing studies solve either the node-wise task or the graph-wise task independently while they are inherently correlated. This work proposes a unified model, AdamGNN, to interactively learn node and graph representations in a mutual-optimisation manner. Compared with existing GNN models and graph pooling methods, AdamGNN enhances the node representation with the learned multi-grained semantics and avoids losing node features and graph structure information during pooling. Specifically, a differentiable pooling operator is proposed to adaptively generate a multi-grained structure that involves meso- and macro-level semantic information in the graph. We also devise the unpooling operator and the flyback aggregator in AdamGNN to better leverage the multi-grained semantics to enhance node representations. The updated node representations can further adjust the graph representation in the next iteration. Experiments on 14 real-world graph datasets show that AdamGNN can significantly outperform 17 competing models on both node- and graph-wise tasks. The ablation studies confirm the effectiveness of AdamGNN's components, and the last empirical analysis further reveals the ingenious ability of AdamGNN in capturing long-range interactions.