Deep Graph Neural Networks with Shallow Subgraph Samplers

While Graph Neural Networks (GNNs) are powerful models for learning representations on graphs, most state-of-the-art models do not have significant accuracy gain beyond two to three layers. Deep GNNs fundamentally need to address: 1). expressivity challenge due to oversmoothing, and 2). computation...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2022-03
Main Authors: Zeng, Hanqing, Zhang, Muhan, Xia, Yinglong, Srivastava, Ajitesh, Malevich, Andrey, Kannan, Rajgopal, Prasanna, Viktor, Long, Jin, Chen, Ren
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Graph neural networks
Graph theory
Graphical representations
Graphs
Machine learning
Model accuracy
Neural networks
Samplers
White noise
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:While Graph Neural Networks (GNNs) are powerful models for learning representations on graphs, most state-of-the-art models do not have significant accuracy gain beyond two to three layers. Deep GNNs fundamentally need to address: 1). expressivity challenge due to oversmoothing, and 2). computation challenge due to neighborhood explosion. We propose a simple "deep GNN, shallow sampler" design principle to improve both the GNN accuracy and efficiency -- to generate representation of a target node, we use a deep GNN to pass messages only within a shallow, localized subgraph. A properly sampled subgraph may exclude irrelevant or even noisy nodes, and still preserve the critical neighbor features and graph structures. The deep GNN then smooths the informative local signals to enhance feature learning, rather than oversmoothing the global graph signals into just "white noise". We theoretically justify why the combination of deep GNNs with shallow samplers yields the best learning performance. We then propose various sampling algorithms and neural architecture extensions to achieve good empirical results. On the largest public graph dataset, ogbn-papers100M, we achieve state-of-the-art accuracy with an order of magnitude reduction in hardware cost.
ISSN:2331-8422