A Large-Scale Database for Graph Representation Learning
arXiv:2011.07682
Abstract
With the rapid emergence of graph representation learning, the construction of new large-scale datasets is necessary to distinguish model capabilities and accurately assess the strengths and weaknesses of each technique. By carefully analyzing existing graph databases, we identify 3 critical components important for advancing the field of graph representation learning: (1) large graphs, (2) many graphs, and (3) class diversity. To date, no single graph database offers all these desired properties. We introduce MalNet, the largest public graph database ever constructed, representing a large-scale ontology of malicious software function call graphs. MalNet contains over 1.2 million graphs, averaging over 15k nodes and 35k edges per graph, across a hierarchy of 47 types and 696 families. Compared to the popular REDDIT-12K database, MalNet offers 105x more graphs, 39x larger graphs on average, and 63x more classes. We provide a detailed analysis of MalNet, discussing its properties and provenance, along with the evaluation of state-of-the-art machine learning and graph neural network techniques. The unprecedented scale and diversity of MalNet offers exciting opportunities to advance the frontiers of graph representation learning--enabling new discoveries and research into imbalanced classification, explainability and the impact of class hardness. The database is publicly available at www.mal-net.org.
arXiv commentPublished in NeurIPS Datasets and Benchmarks Track, 2021
References in corpus (12)
- SMOTE: Synthetic Minority Over-sampling Technique
- Fast Graph Representation Learning with PyTorch Geometric
- Simplifying Graph Convolutional Networks
- Pitfalls of Graph Neural Network Evaluation
- graph2vec: Learning Distributed Representations of Graphs
- Massively Multitask Networks for Drug Discovery
- NetLSD: Hearing the Shape of a Graph
- AndroZoo++: Collecting Millions of Android Apps and Their Metadata for the Research Community
- A simple yet effective baseline for non-attributed graph classification
- A Simple Baseline Algorithm for Graph Classification
- Just SLaQ When You Approximate: Accurate Spectral Distances for Web-Scale Graphs
- ELF: An Early-Exiting Framework for Long-Tailed Classification