Heterophily-Aware Graph Attention Network

TL;DR

This paper introduces HA-GAT, a novel graph neural network that models edge heterophily through a heterophily-aware attention scheme, significantly improving performance on diverse graph datasets.

Contribution

The paper proposes a heterophily-aware attention mechanism and a new GNN architecture, HA-GAT, to better handle heterophilic graphs by modeling edge heterophily explicitly.

Findings

HA-GAT achieves state-of-the-art results on eight datasets.

Modeling edge heterophily improves local attention patterns.

The approach is effective across various homophily ratios.

Abstract

Graph Neural Networks (GNNs) have shown remarkable success in graph representation learning. Unfortunately, current weight assignment schemes in standard GNNs, such as the calculation based on node degrees or pair-wise representations, can hardly be effective in processing the networks with heterophily, in which the connected nodes usually possess different labels or features. Existing heterophilic GNNs tend to ignore the modeling of heterophily of each edge, which is also a vital part in tackling the heterophily problem. In this paper, we firstly propose a heterophily-aware attention scheme and reveal the benefits of modeling the edge heterophily, i.e., if a GNN assigns different weights to edges according to different heterophilic types, it can learn effective local attention patterns, which enable nodes to acquire appropriate information from distinct neighbors. Then, we propose a novel Heterophily-Aware Graph Attention Network (HA-GAT) by fully exploring and utilizing the local distribution as the underlying heterophily, to handle the networks with different homophily ratios. To demonstrate the effectiveness of the proposed HA-GAT, we analyze the proposed heterophily-aware attention scheme and local distribution exploration, by seeking for an interpretation from their mechanism. Extensive results demonstrate that our HA-GAT achieves state-of-the-art performances on eight datasets with different homophily ratios in both the supervised and semi-supervised node classification tasks.