Adaptive Universal Generalized PageRank Graph Neural Network

By Eli Chien, Jianhao Peng, Pan Li, Olgica Milenkovic. International Conference on Learning Representations, 2020

1,039 citations recorded by the source corpus.

In many important graph data processing applications the acquired information includes both node features and observations of the graph topology. Graph neural networks (GNNs) are designed to exploit both sources of evidence but they do not optimally trade-off their utility and integrate them in a manner that is also universal. Here, universality refers to independence on homophily or heterophily graph assumptions. We address these issues by introducing a new Generalized PageRank (GPR) GNN architecture that adaptively learns the GPR weights so as to jointly optimize node feature and topological information extraction, regardless of the extent to which the node labels are homophilic or heterophilic. Learned GPR weights automatically adjust to the node label pattern, irrelevant on the type of initialization, and thereby guarantee excellent learning performance for label patterns that are usually hard to handle. Furthermore, they allow one to avoid feature over-smoothing, a process which renders feature information nondiscriminative, without requiring the network to be shallow. Our accompanying theoretical analysis of the GPR-GNN method is facilitated by novel synthetic benchmark datasets generated by the so-called contextual stochastic block model. We also compare the performance of our GNN architecture with that of several state-of-the-art GNNs on the problem of node-classification, using well-known benchmark homophilic and heterophilic datasets. The results demonstrate that GPR-GNN offers significant performance improvement compared to existing techniques on both synthetic and benchmark data.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/0ee0801ba010a441403f9ed666ef9bf006b3aa07

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

1. Muhan Zhang
2. Xiyuan Wang
3. Jiliang Tang
4. Sergey Levine
5. Yao Ma
6. Kevin Swersky
7. Doina Precup
8. Sitao Luan

Citations in the corpus

• Two Sides of the Same Coin: Heterophily and Oversmoothing in Graph Convolutional Neural Networks (2021, background): “Finally, Chien et al. (2021) recently proposed a pagerank-based model that performs well under heterophily and alleviates the oversmoothing problem.”
• Revisiting Heterophily For Graph Neural Networks (2022, methodology): “In some cases, even simple Multi-Layer Perceptrons (MLPs) can outperform GNNs by a large margin on relational data [46, 29, 32, 8].”
• BernNet: Learning Arbitrary Graph Spectral Filters via Bernstein Approximation (2021, methodology): “As shown in [5] these 5 datasets are homophilic graphs on which the connected nodes tend to share the same label.”
• How Powerful are Spectral Graph Neural Networks (2022, methodology): “By designing/learning the polynomial coefficients, spectral GNNs can simulate a wide range of filters (low-pass, band-pass, high-pass) in the spectral domain, enabling GNNs to work on not only homophilic but also heterophilic graphs (Chien et al., 2021).”
• Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods (2021, neutral): “Non-homophilous methods: H 2 GCN [82], MixHop [1], GPR-GNN [17], GCNII [15], and LINKX (ours).”
• Is Homophily a Necessity for Graph Neural Networks? (2021, methodology): “Due to this neighborhood aggregation mechanism, existing literature posits that strong homophily of the underlying graph is a necessity for GNNs to achieve good performance on SSNC [1, 41, 40, 4, 22, 12, 14, 35].”
• A critical look at the evaluation of GNNs under heterophily: are we really making progress? (2023, methodology): “Heterophily-specific models We use eight models designed for node classification under het-erophily: H 2 GCN (Zhu et al., 2020), CPGNN (Zhu et al., 2021), GPR-GNN (Chien et al., 2021), FSGNN (Maurya et al., 2022), GloGNN (Li et al., 2022), FAGCN (Bo et al., 2021), GBK-8 GNN (Du et al., 2022), and Jaco”
• Finding Global Homophily in Graph Neural Networks When Meeting Heterophily (2022, methodology): “…et al., 2017), MixHop (Abu-El-Haija et al., 2019) and GCNII (Chen et al., 2020); (3) heterophilous-graph-oriented methods: H2GCN (Zhu et al., 2020b), WRGAT (Suresh et al., 2021), GPR-GNN (Chien et al., 2020), GGCN (Yan et al., 2021), ACM-GCN (Luan et al., 2021) and LINKX (Lim et al., 2021).”