Temporal Graph Benchmark for Machine Learning on Temporal Graphs

TL;DR

The paper introduces the Temporal Graph Benchmark (TGB), a comprehensive collection of large-scale, diverse datasets and evaluation protocols for advancing machine learning research on temporal graphs, highlighting variability in model performance and providing tools for reproducible experiments.

Contribution

The paper presents the TGB benchmark, including datasets, evaluation protocols, and an automated pipeline, facilitating standardized, reproducible research on temporal graph machine learning models.

Findings

Model performance varies significantly across datasets.

Simple methods often outperform complex models on dynamic node prediction.

TGB enables reproducible and accessible temporal graph research.

Abstract

We present the Temporal Graph Benchmark (TGB), a collection of challenging and diverse benchmark datasets for realistic, reproducible, and robust evaluation of machine learning models on temporal graphs. TGB datasets are of large scale, spanning years in duration, incorporate both node and edge-level prediction tasks and cover a diverse set of domains including social, trade, transaction, and transportation networks. For both tasks, we design evaluation protocols based on realistic use-cases. We extensively benchmark each dataset and find that the performance of common models can vary drastically across datasets. In addition, on dynamic node property prediction tasks, we show that simple methods often achieve superior performance compared to existing temporal graph models. We believe that these findings open up opportunities for future research on temporal graphs. Finally, TGB provides an automated machine learning pipeline for reproducible and accessible temporal graph research, including data loading, experiment setup and performance evaluation. TGB will be maintained and updated on a regular basis and welcomes community feedback. TGB datasets, data loaders, example codes, evaluation setup, and leaderboards are publicly available at https://tgb.complexdatalab.com/.