Alternatively, the spectral kernels are defined in the spectral domain [2,13,14]. Nonetheless, when modeling graph-structured data, prior kernels face severalchallenges.

Weprovidestrong empirical evidence thatexactinference does not have this pathology, hence it is due to the approximation and not the model.

These authors contributed equally to this work.

Graph Contrastive Learning (GCL) has shown superior performance in representation learning in graph-structured data. Despite their success, most existing GCL methods rely on prefabricated graph augmentation and homophily assumptions. Thus, they fail to generalize well to heterophilic graphs where connected nodes may have different class labels and dissimilar features.

Recently, researchers from convex optimization proposed the notions of "relative Lipschitz continuity" and "relative strong convexity". Both of the notions are generalizations oftheirclassicalcounterparts.