Adaptive Propagation Graph Convolutional Network

Abstract

Graph convolutional networks (GCNs) are a family of neural network models that perform inference on graph data by interleaving vertex-wise operations and message-passing exchanges across nodes. Concerning the latter, two key questions arise: (i) how to design a differentiable exchange protocol (e.g., a 1-hop Laplacian smoothing in the original GCN), and (ii) how to characterize the trade-off in complexity with respect to the local updates. In this paper, we show that state-of-the-art results can be achieved by adapting the number of communication steps independently at every node. In particular, we endow each node with a halting unit (inspired by Graves' adaptive computation time) that after every exchange decides whether to continue communicating or not. We show that the proposed adaptive propagation GCN (AP-GCN) achieves superior or similar results to the best proposed models so far on a number of benchmarks, while requiring a small overhead in terms of additional parameters. We also investigate a regularization term to enforce an explicit trade-off between communication and accuracy. The code for the AP-GCN experiments is released as an open-source library.

Published in IEEE Transaction on Neural Networks and Learning Systems

References in corpus (19)

• Adam: A Method for Stochastic Optimization
• A Comprehensive Survey on Graph Neural Networks
• Neural Message Passing for Quantum Chemistry
• Geometric deep learning: going beyond Euclidean data
• Spectral Networks and Locally Connected Networks on Graphs
• Fast Graph Representation Learning with PyTorch Geometric
• Graph Convolutional Matrix Completion
• Representation Learning on Graphs with Jumping Knowledge Networks
• Graph Neural Networks with convolutional ARMA filters
• Predict then Propagate: Graph Neural Networks meet Personalized PageRank
• Adaptive Computation Time for Recurrent Neural Networks
• A Gentle Introduction to Deep Learning for Graphs
• Kernel-based Reconstruction of Graph Signals
• Diffusion Improves Graph Learning
• Adaptive Least Mean Squares Estimation of Graph Signals
• Modeling Relational Data with Graph Convolutional Networks
• Bridging the Gap between Spatial and Spectral Domains: A Survey on Graph Neural Networks
• Bootstrapped Graph Diffusions: Exposing the Power of Nonlinearity
• Constant Time Graph Neural Networks

Cited by in corpus (16)

• Principal Neighbourhood Aggregation for Graph Nets
• Graph Attention Multi-Layer Perceptron
• FairDrop: Biased Edge Dropout for Enhancing Fairness in Graph Representation Learning
• Beyond Low-pass Filtering: Graph Convolutional Networks with Automatic Filtering
• PaSca: a Graph Neural Architecture Search System under the Scalable Paradigm
• Model Degradation Hinders Deep Graph Neural Networks
• A Meta-Learning Approach for Training Explainable Graph Neural Networks
• ROD: Reception-aware Online Distillation for Sparse Graphs
• How hard is to distinguish graphs with graph neural networks?
• Evaluating Deep Graph Neural Networks
• Node Dependent Local Smoothing for Scalable Graph Learning
• Adversarial Graph Disentanglement
• GMLP: Building Scalable and Flexible Graph Neural Networks with Feature-Message Passing
• Stacked Graph Filter
• EWasteNet: A Two-Stream Data Efficient Image Transformer Approach for E-Waste Classification
• Aerial Images Meet Crowdsourced Trajectories: A New Approach to Robust Road Extraction