Revisiting Heterophily For Graph Neural Networks

By Sitao Luan, Chenqing Hua, Qincheng Lu, Jiaqi Zhu, Mingde Zhao, Shuyuan Zhang, Xiaoming Chang, Doina Precup. Neural Information Processing Systems, 2022

325 citations recorded by the source corpus.

Graph Neural Networks (GNNs) extend basic Neural Networks (NNs) by using graph structures based on the relational inductive bias (homophily assumption). While GNNs have been commonly believed to outperform NNs in real-world tasks, recent work has identified a non-trivial set of datasets where their performance compared to NNs is not satisfactory. Heterophily has been considered the main cause of this empirical observation and numerous works have been put forward to address it. In this paper, we first revisit the widely used homophily metrics and point out that their consideration of only graph-label consistency is a shortcoming. Then, we study heterophily from the perspective of post-aggregation node similarity and define new homophily metrics, which are potentially advantageous compared to existing ones. Based on this investigation, we prove that some harmful cases of heterophily can be effectively addressed by local diversification operation. Then, we propose the Adaptive Channel Mixing (ACM), a framework to adaptively exploit aggregation, diversification and identity channels node-wisely to extract richer localized information for diverse node heterophily situations. ACM is more powerful than the commonly used uni-channel framework for node classification tasks on heterophilic graphs and is easy to be implemented in baseline GNN layers. When evaluated on 10 benchmark node classification tasks, ACM-augmented baselines consistently achieve significant performance gain, exceeding state-of-the-art GNNs on most tasks without incurring significant computational burden.

Subfields: Artificial Intelligence, Materials Chemistry

View Paper: https://www.semanticscholar.org/paper/4becb19c87f0526d9a3a2c15497e0b1c40b576e2

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

A critical look at the evaluation of GNNs under heterophily: are we really making progress? (2023, background): “…were used in numerous papers (Zhu et al., 2021; 2020; Chien et al., 2021; Yan et al., 2022; Maurya et al., 2022; Li et al., 2022; He et al., 2022; Wang & Zhang, 2022; Wang et al., 2022; Du et al., 2022; Suresh et al., 2021; Bo et al., 2021; Ma et al., 2022; Luan et al., 2022; Bodnar et al., 2022).”
Graph Neural Networks for Graphs With Heterophily: A Survey (2022, unknown): “The results indicate that although some early specialized models are occasionally surpassed by well-tuned traditional GNNs, recent state-of-the-art architectures—such as ACM-GCN [59], AGS-GNN [161], and PD-GNN [162]—consistently outperform traditional baselines.”