[Paper Review] GraphSAINT: Graph Sampling Based Inductive Learning Method

Hanqing Zeng, Hongkuan Zhou+3 more|arXiv (Cornell University)|Jul 10, 2019

Advanced Graph Neural NetworksComputer Science33 references336 citations

TL;DR

GraphSAINT presents a graph sampling framework for inductive learning on large graphs by training on sampled subgraphs.

ABSTRACT

Graph Convolutional Networks (GCNs) are powerful models for learning representations of attributed graphs. To scale GCNs to large graphs, state-of-the-art methods use various layer sampling techniques to alleviate the "neighbor explosion" problem during minibatch training. We propose GraphSAINT, a graph sampling based inductive learning method that improves training efficiency and accuracy in a fundamentally different way. By changing perspective, GraphSAINT constructs minibatches by sampling the training graph, rather than the nodes or edges across GCN layers. Each iteration, a complete GCN is built from the properly sampled subgraph. Thus, we ensure fixed number of well-connected nodes in all layers. We further propose normalization technique to eliminate bias, and sampling algorithms for variance reduction. Importantly, we can decouple the sampling from the forward and backward propagation, and extend GraphSAINT with many architecture variants (e.g., graph attention, jumping connection). GraphSAINT demonstrates superior performance in both accuracy and training time on five large graphs, and achieves new state-of-the-art F1 scores for PPI (0.995) and Reddit (0.970).

Motivation & Objective

• Enable scalable inductive learning on large graphs using subgraph sampling.
• Control computational resources via node and edge budgets during sampling.
• Leverage root nodes and random walks to construct informative subgraphs for training.

Proposed method

• Specify sampling parameters: node budget n, edge budget m, number of roots r, and random walk length h.
• Sample a subgraph Gs(Vs, Es) from the original graph G(V,E) to create training data.
• Convert sampled subgraphs into training units suitable for graph neural networks.
• Use root-based or walk-based strategies to drive diverse and representative subgraph samples.

Experimental results

Research questions

• RQ1 How can subgraph sampling affect the accuracy and generalization of inductive graph learning?
• RQ2 What are the trade-offs between node/edge budgets and model performance?
• RQ3 Can random walk-based sampling produce informative subgraphs for training?
• RQ4 How does the sampling strategy impact scalability to very large graphs?

Key findings

• Not available in the provided excerpt; the supplied text does not include experimental results or quantitative findings.