On Addressing the Limitations of Graph Neural Networks

06/22/2023

∙

This report gives a summary of two problems about graph convolutional networks (GCNs): over-smoothing and heterophily challenges, and outlines future directions to explore.

READ FULL TEXT

On Addressing the Limitations of Graph Neural Networks

Over-smoothing Effect of Graph Convolutional Networks

Learning Theory Can (Sometimes) Explain Generalisation in Graph Neural Networks

Fisher-Bures Adversary Graph Convolutional Networks

Graph Convolutional Neural Networks as Parametric CoKleisli morphisms

Predicting Biomedical Interactions with Probabilistic Model Selection for Graph Neural Networks

A Network Science perspective of Graph Convolutional Networks: A survey

Kriging Convolutional Networks

On Addressing the Limitations of Graph Neural Networks

Related Research

Over-smoothing Effect of Graph Convolutional Networks

Learning Theory Can (Sometimes) Explain Generalisation in Graph Neural Networks

Fisher-Bures Adversary Graph Convolutional Networks

Graph Convolutional Neural Networks as Parametric CoKleisli morphisms

Predicting Biomedical Interactions with Probabilistic Model Selection for Graph Neural Networks

A Network Science perspective of Graph Convolutional Networks: A survey

Kriging Convolutional Networks