<ul><li>Recent research has witnessed the progress of Graph Neural Networks (GNNs) in graph data representation.</li><li>GNNs face the challenge of structural imbalance, and existing solutions do not account for graph heterophily.</li><li>The HeRB (Heterophily-Resolved Structure Balancer) method is proposed to address this problem.</li><li>Experimental results show that HeRB outperforms other methods on benchmark datasets.</li></ul>

HeRB: Heterophily-Resolved Structure Balancer for Graph Neural Networks

Discover more