Abstract:Homophily principle, i.e., nodes with the same labels are more likely to be connected, has been believed to be the main reason for the performance superiority of Graph Neural Networks (GNNs) over Neural Networks on node classification tasks. Recent research suggests that, even in the absence of homophily, the advantage of GNNs still exists as long as nodes from the same class share similar neighborhood patterns. However, this argument only considers intra-class Node Distinguishability (ND) but neglects inter-class ND, which provides incomplete understanding of homophily on GNNs. In this paper, we first demonstrate such deficiency with examples and argue that an ideal situation for ND is to have smaller intra-class ND than inter-class ND. To formulate this idea and study ND deeply, we propose Contextual Stochastic Block Model for Homophily (CSBM-H) and define two metrics, Probabilistic Bayes Error (PBE) and negative generalized Jeffreys divergence, to quantify ND. With the metrics, we visualize and analyze how graph filters, node degree distributions and class variances influence ND, and investigate the combined effect of intra- and inter-class ND. Besides, we discovered the mid-homophily pitfall, which occurs widely in graph datasets. Furthermore, we verified that, in real-work tasks, the superiority of GNNs is indeed closely related to both intra- and inter-class ND regardless of homophily levels. Grounded in this observation, we propose a new hypothesis-testing based performance metric beyond homophily, which is non-linear, feature-based and can provide statistical threshold value for GNNs' the superiority. Experiments indicate that it is significantly more effective than the existing homophily metrics on revealing the advantage and disadvantage of graph-aware modes on both synthetic and benchmark real-world datasets.

Improving Node Classification with Neural Tangent Kernel: A Graph Neural Network Approach

Label Propagation across Graphs: Node Classification using Graph Neural Tangent Kernels

Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels

A Novel Two-Channel Classification Approach Using Graph Attention Network with K-Nearest Neighbor

When Do Graph Neural Networks Help with Node Classification: Investigating the Homophily Principle on Node Distinguishability

Neural Tangent Kernels Motivate Graph Neural Networks with Cross-Covariance Graphs

Improving Node Classification Accuracy of GNN through Input and Output Intervention

Bag of Tricks for Node Classification with Graph Neural Networks

New Insights into Graph Convolutional Networks using Neural Tangent Kernels

When Do Graph Neural Networks Help with Node Classification? Investigating the Impact of Homophily Principle on Node Distinguishability

Graph Data Augmentation for Node Classification

Path-enhanced Graph Convolutional Networks for Node Classification Without Features.

A Graph Neural Network-Based Node Classification Model on Class-Imbalanced Graph Data

Graph Neural Networks and Their Current Applications in Bioinformatics

GraphAny: A Foundation Model for Node Classification on Any Graph

Ergcn: Data Enhancement Based Robust Graph Convolutional Network for Node Classification

GraphQNTK: Quantum Neural Tangent Kernel for Graph Data

Graph Neural Networks Inspired by Classical Iterative Algorithms

Neural Trees for Learning on Graphs

Improving Node Classification by Co-training Node Pair Classification: A Novel Training Framework for General Graph Neural Networks

Improving Your Graph Neural Networks: A High-Frequency Booster