GLANCE: Graph Logic Attention Network with Cluster Enhancement for Heterophilous Graph Representation Learning

TL;DR

GLANCE introduces a novel graph neural network framework that enhances heterophilous graph learning through logic-guided reasoning, dynamic graph refinement, and clustering, leading to improved interpretability and robustness.

Contribution

The paper presents GLANCE, a new GNN framework that integrates logic reasoning, attention-based edge pruning, and clustering to better handle heterophilous graphs.

Findings

GLANCE achieves competitive performance on benchmark heterophilous datasets.

The framework provides interpretable and robust graph representations.

Experimental results demonstrate effectiveness over existing methods.

Abstract

Graph Neural Networks (GNNs) have demonstrated significant success in learning from graph-structured data but often struggle on heterophilous graphs, where connected nodes differ in features or class labels. This limitation arises from indiscriminate neighbor aggregation and insufficient incorporation of higher-order structural patterns. To address these challenges, we propose GLANCE (Graph Logic Attention Network with Cluster Enhancement), a novel framework that integrates logic-guided reasoning, dynamic graph refinement, and adaptive clustering to enhance graph representation learning. GLANCE combines a logic layer for interpretable and structured embeddings, multi-head attention-based edge pruning for denoising graph structures, and clustering mechanisms for capturing global patterns. Experimental results in benchmark datasets, including Cornell, Texas, and Wisconsin, demonstrate that GLANCE achieves competitive performance, offering robust and interpretable solutions for heterophilous graph scenarios. The proposed framework is lightweight, adaptable, and uniquely suited to the challenges of heterophilous graphs.