Bayesian Semi-supervised Learning with Graph Gaussian Processes

Abstract

We propose a data-efficient Gaussian process-based Bayesian approach to the semi-supervised learning problem on graphs. The proposed model shows extremely competitive performance when compared to the state-of-the-art graph neural networks on semi-supervised learning benchmark experiments, and outperforms the neural networks in active learning experiments where labels are scarce. Furthermore, the model does not require a validation data set for early stopping to control over-fitting. Our model can be viewed as an instance of empirical distribution regression weighted locally by network connectivity. We further motivate the intuitive construction of the model with a Bayesian linear model interpretation where the node features are filtered by an operator related to the graph Laplacian. The method can be easily implemented by adapting off-the-shelf scalable variational inference algorithms for Gaussian processes.

To appear in NIPS 2018 Fixed an error in Figure 2. The previous arxiv version contains two identical sub-figures

References in corpus (9)

• The Emerging Field of Signal Processing on Graphs: Extending High-Dimensional Data Analysis to Networks and Other Irregular Domains
• Geometric deep learning: going beyond Euclidean data
• Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering
• Diffusion-Convolutional Neural Networks
• Gated Graph Sequence Neural Networks
• Revisiting Semi-Supervised Learning with Graph Embeddings
• Additive Gaussian Processes
• MCMC for Variationally Sparse Gaussian Processes
• Graph-Based Active Learning: A New Look at Expected Error Minimization

Cited by in corpus (6)

• Graph Neural Networks: A Review of Methods and Applications
• Relation Modeling with Graph Convolutional Networks for Facial Action Unit Detection
• A Flexible Generative Framework for Graph-based Semi-supervised Learning
• Infinitely Wide Graph Convolutional Networks: Semi-supervised Learning via Gaussian Processes
• Training on Test Data with Bayesian Adaptation for Covariate Shift
• Investigating the Effect of Intraclass Variability in Temporal Ensembling