TUDataset: A collection of benchmark datasets for learning with graphs

Abstract

Recently, there has been an increasing interest in (supervised) learning with graph data, especially using graph neural networks. However, the development of meaningful benchmark datasets and standardized evaluation procedures is lagging, consequently hindering advancements in this area. To address this, we introduce the TUDataset for graph classification and regression. The collection consists of over 120 datasets of varying sizes from a wide range of applications. We provide Python-based data loaders, kernel and graph neural network baseline implementations, and evaluation tools. Here, we give an overview of the datasets, standardized evaluation procedures, and provide baseline experiments. All datasets are available at www.graphlearning.io. The experiments are fully reproducible from the code available at www.github.com/chrsmrrs/tudataset.

ICML 2020 workshop "Graph Representation Learning and Beyond"

References in corpus (4)

• Fast Graph Representation Learning with PyTorch Geometric
• What's in a crowd? Analysis of face-to-face behavioral networks
• Alchemy: A Quantum Chemistry Dataset for Benchmarking AI Models
• IPC: A Benchmark Data Set for Learning with Graph-Structured Data

Cited by in corpus (7)

• A Meta-Learning Approach for Graph Representation Learning in Multi-Task Settings
• ADOM: Accelerated Decentralized Optimization Method for Time-Varying Networks
• Should Graph Neural Networks Use Features, Edges, Or Both?
• Learning Graphons via Structured Gromov-Wasserstein Barycenters
• Diversified Multiscale Graph Learning with Graph Self-Correction
• Improving the Expressive Power of Graph Neural Network with Tinhofer Algorithm
• Scaling up graph homomorphism for classification via sampling