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

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Sun, Zhongtian, Harit, Anoushka, Cristea, Alexandra, Donnelly, Christl A., Liò, Pietro
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916975591030784
author Sun, Zhongtian
Harit, Anoushka
Cristea, Alexandra
Donnelly, Christl A.
Liò, Pietro
author_facet Sun, Zhongtian
Harit, Anoushka
Cristea, Alexandra
Donnelly, Christl A.
Liò, Pietro
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2507_18521
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GLANCE: Graph Logic Attention Network with Cluster Enhancement for Heterophilous Graph Representation Learning
Sun, Zhongtian
Harit, Anoushka
Cristea, Alexandra
Donnelly, Christl A.
Liò, Pietro
Machine Learning
Artificial Intelligence
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.
title GLANCE: Graph Logic Attention Network with Cluster Enhancement for Heterophilous Graph Representation Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2507.18521