TUDataset: A collection of benchmark datasets for learning with graphs

By Christopher Morris, Nils M. Kriege, Franka Bause, Kristian Kersting, Petra Mutzel, Marion Neumann. arXiv.org, 2020

1,113 citations recorded by the source corpus.

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 this http URL. The experiments are fully reproducible from the code available at this http URL.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/3302081b8878de505b69900e22b8137c5d931be7

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Graph Contrastive Learning Automated (2021, methodology): “The results of semi-supervised learning & unsupervised representation learning on TUDataset are in Tables 4 & 5, respectively.”
Adversarial Graph Augmentation to Improve Graph Contrastive Learning (2021, methodology): “We use datasets from Open Graph Benchmark (OGB) [52], TU Dataset [73] and ZINC [74] for graph-level property classiﬁcation and regression.”
Graph Retrieval-Augmented Generation: A Survey (2024, neutral): “One typical kind of TAGs is Knowledge Graphs (KGs), where nodes are entities, edges are relations among entities, and text attributes are the names of entities and relations.”
Self-Supervised Learning of Graph Neural Networks: A Unified Review (2021, methodology): “Datasets for chemicalmolecular property prediction are also categorized as bioinfomatics datasets in TUDataset.”
Graph Neural Networks with Learnable Structural and Positional Representations (2021, methodology): “IDBM-BINARY and IMDB-MULTI contain 1,000 and 1,500 graphs respectively which are egonetworks extracted from actor collaboration graphs.”
Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks (2021, methodology): “The deep theoretical link between the WL graph isomorphism test and message passing GNNs are well known (Xu et al., 2019b).”
Contrastive Learning with Hard Negative Samples (2020, methodology): “Each embedding is evaluated using the average accuracy 10-fold cross-validation using an SVM as the classiﬁer (in line with the approach taken by Morris et al. (2020)).”
SimGRACE: A Simple Framework for Graph Contrastive Learning without Data Augmentation (2022, methodology): “For unsupervised and semi-supervised learning, we use datasets from the benchmark TUDataset [27], including graph data for various social networks [2, 52] and biochemical molecules [9, 32].”