Preprint
Aug 2026
Disentangling Homophily and Rarity: Explaining Failure in Graph Neural Networks
It is demonstrated that the information needed to classify heterophilic nodes correctly is often recoverable by retraining the classification head of a model, or even just the final linear classification layer.
Preben M. Ness, Fariz Ikhwantri, D. Marijan
· 0 citations