PathSAGE: Spatial Graph Attention Neural Networks With Random Path Sampling

TL;DR

PathSAGE introduces a novel graph neural network that uses random path sampling and Transformer encoding to effectively learn high-order topological features while avoiding common deep GCN issues.

Contribution

The paper presents PathSAGE, a single-layer model that samples random paths and employs Transformer encoders to enhance performance on non-Euclidean data.

Findings

Achieves comparable performance with state-of-the-art models

Effectively captures high-order topological information

Avoids problems like over-smoothing and neighbor explosion

Abstract

Graph Convolutional Networks (GCNs) achieve great success in non-Euclidean structure data processing recently. In existing studies, deeper layers are used in CCNs to extract deeper features of Euclidean structure data. However, for non-Euclidean structure data, too deep GCNs will confront with problems like "neighbor explosion" and "over-smoothing", it also cannot be applied to large datasets. To address these problems, we propose a model called PathSAGE, which can learn high-order topological information and improve the model's performance by expanding the receptive field. The model randomly samples paths starting from the central node and aggregates them by Transformer encoder. PathSAGE has only one layer of structure to aggregate nodes which avoid those problems above. The results of evaluation shows that our model achieves comparable performance with the state-of-the-art models in inductive learning tasks.