SGAT: Simplicial Graph Attention Network

TL;DR

SGAT introduces a simplicial complex-based graph neural network that captures high-order interactions in heterogeneous graphs, leading to improved node classification performance over existing methods.

Contribution

The paper proposes SGAT, a novel approach using simplicial complexes and attention mechanisms to model complex high-order interactions in heterogeneous graphs.

Findings

SGAT outperforms state-of-the-art methods in node classification.

SGAT effectively captures structural information with random node features.

Numerical experiments validate the superiority of SGAT.

Abstract

Heterogeneous graphs have multiple node and edge types and are semantically richer than homogeneous graphs. To learn such complex semantics, many graph neural network approaches for heterogeneous graphs use metapaths to capture multi-hop interactions between nodes. Typically, features from non-target nodes are not incorporated into the learning procedure. However, there can be nonlinear, high-order interactions involving multiple nodes or edges. In this paper, we present Simplicial Graph Attention Network (SGAT), a simplicial complex approach to represent such high-order interactions by placing features from non-target nodes on the simplices. We then use attention mechanisms and upper adjacencies to generate representations. We empirically demonstrate the efficacy of our approach with node classification tasks on heterogeneous graph datasets and further show SGAT's ability in extracting structural information by employing random node features. Numerical experiments indicate that SGAT performs better than other current state-of-the-art heterogeneous graph learning methods.