Towards Sparse Hierarchical Graph Classifiers

Abstract

Recent advances in representation learning on graphs, mainly leveraging graph convolutional networks, have brought a substantial improvement on many graph-based benchmark tasks. While novel approaches to learning node embeddings are highly suitable for node classification and link prediction, their application to graph classification (predicting a single label for the entire graph) remains mostly rudimentary, typically using a single global pooling step to aggregate node features or a hand-designed, fixed heuristic for hierarchical coarsening of the graph structure. An important step towards ameliorating this is differentiable graph coarsening---the ability to reduce the size of the graph in an adaptive, data-dependent manner within a graph neural network pipeline, analogous to image downsampling within CNNs. However, the previous prominent approach to pooling has quadratic memory requirements during training and is therefore not scalable to large graphs. Here we combine several recent advances in graph neural network design to demonstrate that competitive hierarchical graph classification results are possible without sacrificing sparsity. Our results are verified on several established graph classification benchmarks, and highlight an important direction for future research in graph-based neural networks.

arXiv commentTo appear in the Workshop on Relational Representation Learning (R2L) at NIPS 2018. 6 pages, 3 figures

References in corpus (11)

• Adam: A Method for Stochastic Optimization
• Semi-Supervised Classification with Graph Convolutional Networks
• Inductive Representation Learning on Large Graphs
• Neural Message Passing for Quantum Chemistry
• Spectral Networks and Locally Connected Networks on Graphs
• Representation Learning on Graphs with Jumping Knowledge Networks
• Learning Convolutional Neural Networks for Graphs
• Discriminative Embeddings of Latent Variable Models for Structured Data
• Gated Graph Sequence Neural Networks
• How Powerful are Graph Neural Networks?
• Flexible Neural Representation for Physics Prediction

Cited by in corpus (46)

• Fast Graph Representation Learning with PyTorch Geometric
• MeshCNN: A Network with an Edge
• Self-Attention Graph Pooling
• Graph Neural Networks: Taxonomy, Advances and Trends
• Understanding Pooling in Graph Neural Networks
• Edge Contraction Pooling for Graph Neural Networks
• Spectral Clustering with Graph Neural Networks for Graph Pooling
• Accurate Learning of Graph Representations with Graph Multiset Pooling
• Graph Convolutional Neural Networks for Body Force Prediction
• Hierarchical Representation Learning in Graph Neural Networks with Node Decimation Pooling
• Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks
• Bridging the Gap Between Spectral and Spatial Domains in Graph Neural Networks
• Path Integral Based Convolution and Pooling for Graph Neural Networks
• Incorporating Symbolic Domain Knowledge into Graph Neural Networks
• Just Jump: Dynamic Neighborhood Aggregation in Graph Neural Networks
• SUGAR: Subgraph Neural Network with Reinforcement Pooling and Self-Supervised Mutual Information Mechanism
• Spectral Multigraph Networks for Discovering and Fusing Relationships in Molecules
• Fake News Detection in Social Media using Graph Neural Networks and NLP Techniques: A COVID-19 Use-case
• Cell Attention Networks
• Inclusion of Domain-Knowledge into GNNs using Mode-Directed Inverse Entailment
• Soft-mask: Adaptive Substructure Extractions for Graph Neural Networks
• Single Image 3D Hand Reconstruction with Mesh Convolutions
• Graph Neural Network for Interpreting Task-fMRI Biomarkers
• Graph Coarsening with Neural Networks
• Deep Representation Learning for Social Network Analysis
• How Framelets Enhance Graph Neural Networks
• The Impact of Global Structural Information in Graph Neural Networks Applications
• MathNet: Haar-Like Wavelet Multiresolution-Analysis for Graph Representation and Learning
• Multi-Channel Graph Convolutional Networks
• Utilising Graph Machine Learning within Drug Discovery and Development
• Generalizing Downsampling from Regular Data to Graphs
• Quaternion Graph Neural Networks
• A pipeline for fair comparison of graph neural networks in node classification tasks
• Set Representation Learning with Generalized Sliced-Wasserstein Embeddings
• Human Action Recognition with Multi-Laplacian Graph Convolutional Networks
• Can neural networks learn persistent homology features?
• Should Graph Neural Networks Use Features, Edges, Or Both?
• Embedding Graphs on Grassmann Manifold
• Unsupervised Learning of Graph Hierarchical Abstractions with Differentiable Coarsening and Optimal Transport
• Multi-Level Attention Pooling for Graph Neural Networks: Unifying Graph Representations with Multiple Localities
• Sparse hierarchical representation learning on molecular graphs
• Heterogeneous Grid Convolution for Adaptive, Efficient, and Controllable Computation
• A Structural Feature-Based Approach for Comprehensive Graph Classification
• Structural Landmarking and Interaction Modelling: on Resolution Dilemmas in Graph Classification
• Graph Neural Networks for Inconsistent Cluster Detection in Incremental Entity Resolution
• Automatic design of novel potential 3CL${}^{\text{pro}}$ and PL${}^{\text{pro}}$ inhibitors