On Performance Discrepancies Across Local Homophily Levels in Graph Neural Networks
Donald Loveland University of Michigan, Ann Arbor Zhu University of Michigan, Ann Arbor Heimann Lawrence Livermore National Lab Fish University of Michigan, Ann Arbor T. Schaub RWTH Aachen University Koutra University of Michigan, Ann Arbor
Abstract
Graph Neural Network (GNN) research has highlighted a relationship between high homophily (i.e., the tendency of nodes of the same class to connect) and strong predictive performance in node classification. However, recent work has found the relationship to be more nuanced, demonstrating that simple GNNs can learn in certain heterophilous settings. To resolve these conflicting findings and align closer to real-world datasets, we go beyond the assumption of a global graph homophily level and study the performance of GNNs when the local homophily level of a node deviates from the global homophily level. Through theoretical and empirical analysis, we systematically demonstrate how shifts in local homophily can introduce performance degradation, leading to performance discrepancies across local homophily levels. We ground the practical implications of this work through granular analysis on five real-world datasets with varying global homophily levels, demonstrating that (a) GNNs can fail to generalize to test nodes that deviate from the global homophily of a graph, and (b) high local homophily does not necessarily confer high performance for a node. We further show that GNNs designed f
中文速览
图神经网络（GNN）在节点分类任务中的表现长期被认为与图的同质性（homophily，即同类节点倾向于相互连接）密切相关，但现有研究大多只关注全局同质性水平，忽视了单个节点的局部同质性可能与全局水平存在显著差异这一现实。这篇论文聚焦于"局部同质性偏移"问题，通过理论推导和扰动分析，严格证明了当测试节点的局部同质性与训练时所依赖的全局同质性出现偏差时，GNN的预测性能会系统性下降，且高局部同质性并不必然带来高预测精度。研究者还设计了一个可精细控制局部同质性分布的合成图生成器，并在五个真实数据集和九种GNN架构上进行了大规模实验验证，发现专为异质图（heterophilous graph）设计的GNN能够在不同局部同质性水平的节点上保持更均匀的表现，从而缩小性能差距。这项工作揭示了GNN的一个全新失效点——对全局同质性的过度依赖——并指出这种结构性不均衡在涉及人的真实场景中可能引发公平性问题，为理解和改进GNN提供了重要的新视角。
原文 arXiv:2306.05557;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2306.05557v4