New Benchmarks for Learning on Non-Homophilous Graphs

Abstract

Much data with graph structures satisfy the principle of homophily, meaning that connected nodes tend to be similar with respect to a specific attribute. As such, ubiquitous datasets for graph machine learning tasks have generally been highly homophilous, rewarding methods that leverage homophily as an inductive bias. Recent work has pointed out this particular focus, as new non-homophilous datasets have been introduced and graph representation learning models better suited for low-homophily settings have been developed. However, these datasets are small and poorly suited to truly testing the effectiveness of new methods in non-homophilous settings. We present a series of improved graph datasets with node label relationships that do not satisfy the homophily principle. Along with this, we introduce a new measure of the presence or absence of homophily that is better suited than existing measures in different regimes. We benchmark a range of simple methods and graph neural networks across our proposed datasets, drawing new insights for further research. Data and codes can be found at https://github.com/CUAI/Non-Homophily-Benchmarks.

arXiv commentIn Workshop on Graph Learning Benchmarks (GLB 2021) at WWW 2021. 10 pages

References in corpus (27)

• Adam: A Method for Stochastic Optimization
• Emergence of scaling in random networks
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Decoupled Weight Decay Regularization
• Semi-Supervised Classification with Graph Convolutional Networks
• Inductive Representation Learning on Large Graphs
• Mixing patterns in networks
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Simplifying Graph Convolutional Networks
• Representation Learning on Graphs with Jumping Knowledge Networks
• Social Structure of Facebook Networks
• Pitfalls of Graph Neural Network Evaluation
• Enhancing Graph Neural Network-based Fraud Detectors against Camouflaged Fraudsters
• Open Graph Benchmark: Datasets for Machine Learning on Graphs
• Simple and Deep Graph Convolutional Networks
• Scaling Graph Neural Networks with Approximate PageRank
• MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixing
• Revisiting Graph Neural Networks: All We Have is Low-Pass Filters
• Benchmarking Graph Neural Networks
• Geom-GCN: Geometric Graph Convolutional Networks
• How to Find Your Friendly Neighborhood: Graph Attention Design with Self-Supervision
• Combining Label Propagation and Simple Models Out-performs Graph Neural Networks
• Residual Correlation in Graph Neural Network Regression
• Contextual Stochastic Block Models
• Non-Local Graph Neural Networks
• Two Sides of the Same Coin: Heterophily and Oversmoothing in Graph Convolutional Neural Networks
• Semi-Supervised Learning with Heterophily

Cited by in corpus (8)

• Beyond Low-pass Filtering: Graph Convolutional Networks with Automatic Filtering
• Is Heterophily A Real Nightmare For Graph Neural Networks To Do Node Classification?
• Graph Neural Networks with Diverse Spectral Filtering
• Spatial Heterophily Aware Graph Neural Networks
• GraphGLOW: Universal and Generalizable Structure Learning for Graph Neural Networks
• You are AllSet: A Multiset Function Framework for Hypergraph Neural Networks
• Node Feature Extraction by Self-Supervised Multi-scale Neighborhood Prediction
• Graph Neural Networks with Feature and Structure Aware Random Walk