Wisconsin

The 'Wisconsin' dataset is a benchmark used to evaluate node-level representation learning methods on heterophilic graphs, specifically assessing their ability to capture structural equivalence and hidden relationships among nodes.

Papers using Wisconsin (2)

GMLM: Bridging Graph Neural Networks and Language Models for Heterophilic Node Classification2025 · 1 citesGLANCE: Graph Logic Attention Network with Cluster Enhancement for Heterophilous Graph Representation Learning2025