Non-separable Spatio-temporal Graph Kernels via SPDEs
arXiv:2111.08524
Abstract
Gaussian processes (GPs) provide a principled and direct approach for inference and learning on graphs. However, the lack of justified graph kernels for spatio-temporal modelling has held back their use in graph problems. We leverage an explicit link between stochastic partial differential equations (SPDEs) and GPs on graphs, introduce a framework for deriving graph kernels via SPDEs, and derive non-separable spatio-temporal graph kernels that capture interaction across space and time. We formulate the graph kernels for the stochastic heat equation and wave equation. We show that by providing novel tools for spatio-temporal GP modelling on graphs, we outperform pre-existing graph kernels in real-world applications that feature diffusion, oscillation, and other complicated interactions.
References in corpus (7)
- T-GCN: A Temporal Graph ConvolutionalNetwork for Traffic Prediction
- Graph Neural Network for Traffic Forecasting: A Survey
- Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting
- Kernel-based Reconstruction of Graph Signals
- Machine Learning on Graphs: A Model and Comprehensive Taxonomy
- Matérn Gaussian processes on Riemannian manifolds
- Chickenpox Cases in Hungary: a Benchmark Dataset for Spatiotemporal Signal Processing with Graph Neural Networks