Revisiting Heterophily in Graph Convolution Networks by Learning Representations Across Topological and Feature Spaces

Open in new window