Graph neural networks for graphs with heterophily: A survey

X Zheng, Y Wang, Y Liu, M Li, M Zhang, D Jin… - arXiv preprint arXiv …, 2022 - arxiv.org
Recent years have witnessed fast developments of graph neural networks (GNNs) that have
benefited myriads of graph analytic tasks and applications. In general, most GNNs depend …

Revisiting heterophily for graph neural networks

S Luan, C Hua, Q Lu, J Zhu, M Zhao… - Advances in neural …, 2022 - proceedings.neurips.cc
Abstract Graph Neural Networks (GNNs) extend basic Neural Networks (NNs) by using
graph structures based on the relational inductive bias (homophily assumption). While …

A critical look at the evaluation of GNNs under heterophily: Are we really making progress?

O Platonov, D Kuznedelev, M Diskin… - arXiv preprint arXiv …, 2023 - arxiv.org
Node classification is a classical graph representation learning task on which Graph Neural
Networks (GNNs) have recently achieved strong results. However, it is often believed that …

Two sides of the same coin: Heterophily and oversmoothing in graph convolutional neural networks

Y Yan, M Hashemi, K Swersky, Y Yang… - … Conference on Data …, 2022 - ieeexplore.ieee.org
In node classification tasks, graph convolutional neural networks (GCNs) have
demonstrated competitive performance over traditional methods on diverse graph data …

Demystifying structural disparity in graph neural networks: Can one size fit all?

H Mao, Z Chen, W Jin, H Han, Y Ma… - Advances in neural …, 2024 - proceedings.neurips.cc
Abstract Recent studies on Graph Neural Networks (GNNs) provide both empirical and
theoretical evidence supporting their effectiveness in capturing structural patterns on both …

When do graph neural networks help with node classification? investigating the homophily principle on node distinguishability

S Luan, C Hua, M Xu, Q Lu, J Zhu… - Advances in …, 2024 - proceedings.neurips.cc
Homophily principle, ie, 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 …

A unified view on graph neural networks as graph signal denoising

Y Ma, X Liu, T Zhao, Y Liu, J Tang, N Shah - Proceedings of the 30th …, 2021 - dl.acm.org
Graph Neural Networks (GNNs) have risen to prominence in learning representations for
graph structured data. A single GNN layer typically consists of a feature transformation and a …

Addressing heterophily in graph anomaly detection: A perspective of graph spectrum

Y Gao, X Wang, X He, Z Liu, H Feng… - Proceedings of the ACM …, 2023 - dl.acm.org
Graph anomaly detection (GAD) suffers from heterophily—abnormal nodes are sparse so
that they are connected to vast normal nodes. The current solutions upon Graph Neural …

Beyond homophily: Reconstructing structure for graph-agnostic clustering

E Pan, Z Kang - International Conference on Machine …, 2023 - proceedings.mlr.press
Graph neural networks (GNNs) based methods have achieved impressive performance on
node clustering task. However, they are designed on the homophilic assumption of graph …

Gbk-gnn: Gated bi-kernel graph neural networks for modeling both homophily and heterophily

L Du, X Shi, Q Fu, X Ma, H Liu, S Han… - Proceedings of the ACM …, 2022 - dl.acm.org
Graph Neural Networks (GNNs) are widely used on a variety of graph-based machine
learning tasks. For node-level tasks, GNNs have strong power to model the homophily …