Hierarchical Multi-Scale Graph Neural Networks: Scalable Heterophilous Learning with Oversmoothing and Oversquashing Mitigation

TL;DR

This paper introduces HMH, a spectral GNN framework that effectively handles heterophilous graphs by mitigating oversmoothing and oversquashing through a hierarchical, multi-scale approach with spectral filtering.

Contribution

The paper proposes a novel spectral graph-learning framework, HMH, that scales efficiently and improves heterophilous graph classification by addressing hub aggregation and signal bottleneck issues.

Findings

HMH outperforms state-of-the-art spectral methods in node and graph classification.

Achieves up to 3% improvement on node classification datasets.

Maintains near-linear scalability in large graph settings.

Abstract

Graphs with heterophily, where adjacent nodes carry different labels, are prevalent in real-world applications, from social networks to molecular interactions. However, existing spectral Graph Neural Network (GNN) approaches tailored for heterophilous graph classification suffer from hub-dominated (node with large degree) aggregation and oversmoothing, as their suboptimal polynomial filters introduce approximation errors and blend distant signals. To address the degree-biased aggregation and suboptimal polynomial filtering, we introduce a Hierarchical Multi-view HAAR (HMH), a novel spectral graph-learning framework that scales in near-linear time . HMH first learns feature- and structure-aware signed affinities via a heterophily-aware encoder, then constructs a soft graph hierarchy guided by these embeddings. At each hierarchical level, HMH constructs a sparse, orthonormal, and locality-aware Haar basis to apply learnable spectral filters in the frequency domain. Finally, skip-connection unpooling layers combine outputs from all hierarchical levels back into the original graph, effectively preventing hub domination and long-range signal bottleneck (over-squashing). Experimentation shows that HMH outperforms state-of-the-art spectral baselines, achieving up to a 3% improvement on node classification and 7% points on graph classification datasets, all while maintaining linear scalability.