OGB-LSC: A Large-Scale Challenge for Machine Learning on Graphs

Abstract

Enabling effective and efficient machine learning (ML) over large-scale graph data (e.g., graphs with billions of edges) can have a great impact on both industrial and scientific applications. However, existing efforts to advance large-scale graph ML have been largely limited by the lack of a suitable public benchmark. Here we present OGB Large-Scale Challenge (OGB-LSC), a collection of three real-world datasets for facilitating the advancements in large-scale graph ML. The OGB-LSC datasets are orders of magnitude larger than existing ones, covering three core graph learning tasks -- link prediction, graph regression, and node classification. Furthermore, we provide dedicated baseline experiments, scaling up expressive graph ML models to the massive datasets. We show that expressive models significantly outperform simple scalable baselines, indicating an opportunity for dedicated efforts to further improve graph ML at scale. Moreover, OGB-LSC datasets were deployed at ACM KDD Cup 2021 and attracted more than 500 team registrations globally, during which significant performance improvements were made by a variety of innovative techniques. We summarize the common techniques used by the winning solutions and highlight the current best practices in large-scale graph ML. Finally, we describe how we have updated the datasets after the KDD Cup to further facilitate research advances. The OGB-LSC datasets, baseline code, and all the information about the KDD Cup are available at https://ogb.stanford.edu/docs/lsc/ .

arXiv commentKDD Cup 2021. NeurIPS Datasets and Benchmarks Track 2021

References in corpus (20)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Neural Message Passing for Quantum Chemistry
• Language Models are Few-Shot Learners
• Cluster-GCN: An Efficient Algorithm for Training Deep and Large Graph Convolutional Networks
• Fast Graph Representation Learning with PyTorch Geometric
• Recent developments in the PySCF program package
• Simplifying Graph Convolutional Networks
• The Open Catalyst 2020 (OC20) Dataset and Community Challenges
• The Anatomy of the Facebook Social Graph
• RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space
• MPNet: Masked and Permuted Pre-training for Language Understanding
• FastGCN: Fast Learning with Graph Convolutional Networks via Importance Sampling
• TUDataset: A collection of benchmark datasets for learning with graphs
• SIGN: Scalable Inception Graph Neural Networks
• Strategies for Pre-training Graph Neural Networks
• Combining Label Propagation and Simple Models Out-performs Graph Neural Networks
• Masked Label Prediction: Unified Message Passing Model for Semi-Supervised Classification
• ForceNet: A Graph Neural Network for Large-Scale Quantum Calculations
• Large-scale graph representation learning with very deep GNNs and self-supervision
• NOTE: Solution for KDD-CUP 2021 WikiKG90M-LSC

Cited by in corpus (24)

• Do Transformers Really Perform Bad for Graph Representation?
• Neural Bellman-Ford Networks: A General Graph Neural Network Framework for Link Prediction
• Graph Attention Multi-Layer Perceptron
• Trustworthy Graph Neural Networks: Aspects, Methods and Trends
• Global Self-Attention as a Replacement for Graph Convolution
• Construction of Knowledge Graphs: State and Challenges
• The Evolution of Distributed Systems for Graph Neural Networks and their Origin in Graph Processing and Deep Learning: A Survey
• IGB: Addressing The Gaps In Labeling, Features, Heterogeneity, and Size of Public Graph Datasets for Deep Learning Research
• Large-Scale Representation Learning on Graphs via Bootstrapping
• Bag of Tricks for Node Classification with Graph Neural Networks
• An end-to-end attention-based approach for learning on graphs
• View-based Explanations for Graph Neural Networks
• Large-scale graph representation learning with very deep GNNs and self-supervision
• Multiresolution Graph Transformers and Wavelet Positional Encoding for Learning Hierarchical Structures
• Molecule3D: A Benchmark for Predicting 3D Geometries from Molecular Graphs
• Why Propagate Alone? Parallel Use of Labels and Features on Graphs
• Fast Quantum Property Prediction via Deeper 2D and 3D Graph Networks
• An Empirical Study of Graph Contrastive Learning
• NOTE: Solution for KDD-CUP 2021 WikiKG90M-LSC
• Learn Locally, Correct Globally: A Distributed Algorithm for Training Graph Neural Networks
• On Graph Neural Network Ensembles for Large-Scale Molecular Property Prediction
• Scaling up graph homomorphism for classification via sampling
• Technical Report of Team GraphMIRAcles in the WikiKG90M-LSC Track of OGB-LSC @ KDD Cup 2021
• Contributions to Representation Learning with Graph Autoencoders and Applications to Music Recommendation