oversmoothing

The paper introduces a common layer equation that unifies graph neural networks into seven components, enabling architectural comparison, theoretical analysis, and insights into issues like oversmoothing and expressivity.

This paper introduces HouseGNN, a graph neural network that uses Householder reflections and GroupSort to address oversmoothing, proving that each layer preserves node-wise Euclidean norm and showing improved behavior at depth.

This paper studies oversmoothing in hypergraph neural networks from a dynamical-systems perspective, proposing a reaction-diffusion framework (HNRD) that preserves node-discriminative variation and achieves depth-robust propagation. Experiments show consistent improvement over baselines.

This paper analyzes oversmoothing in Neural Sheaf Diffusion (NSD) as a representation degeneracy phenomenon using quiver theory and Geometric Invariant Theory. It proposes moment-map-inspired regularizers and explores non-uniform stalk dimensions to mitigate this issue in heterophilic graph benchmarks.

This paper introduces HMH, a hierarchical multi-scale Graph Neural Network framework designed to address oversmoothing and oversquashing in heterophilous graphs. It utilizes spectral filters with Haar bases to achieve scalable learning and improved performance on node and graph classification tasks.