AdaKernel: Learning Adaptive Kernel Parameters for Spatiotemporal Graph Neural Networks
For researchers in spatiotemporal data analysis, AdaKernel provides a principled way to adapt kernel parameters, improving model accuracy without discarding geometric structure.
AdaKernel learns adaptive kernel parameters for spatiotemporal GNNs, theoretically proving that misspecified parameters cause approximation errors. It consistently improves multiple GNN architectures on Kriging, Imputation, and Forecasting tasks, outperforming both fixed priors and fully latent graph structures.
Modeling spatial dependencies is central to spatiotemporal data analysis using Graph Neural Networks (GNNs). Traditional methods rely on distance-based kernels with predefined parameters, which restricts model capacity. Although generic adaptive mechanisms (e.g., Graph Attention Networks) offer flexibility, they often fail to capture the underlying geometric structure, performing worse than distance-based models in data-sparse scenarios. Addressing this, we revisit the kernel parameterization problem and theoretically prove that misspecified kernel parameters introduce unavoidable approximation errors in GNNs. To overcome this, we propose AdaKernel, a simple yet effective approach that learns adaptive kernel parameters within the neural network. Unlike methods that learn graph structures from scratch, AdaKernel adopts a structure-preserving strategy that optimizes the scale of physical interactions rather than discarding them. Extensive experiments on Kriging, Imputation, and Forecasting demonstrate that AdaKernel consistently improves various GNN architectures and outperforms model-agnostic adaptive baselines, validating that accurately learned kernel parameters are superior to both fixed priors and fully latent graph structures.