Yue Sun ¹ Chao Chen ¹ Yuesheng Xu ^1* Sihong Xie ² Rick S. Blum ^1* Parv Venkitasubramaniam ^1*

• ¹ Lehigh University, Bethlehem, United States
• ² Hongkong University of Science and Technology(Guangzhou), Guangzhou, China

Graph Neural Networks (GNNs) have gained significant attention in diverse domains, ranging from urban planning to pandemic management. Ensuring both accuracy and robustness in GNNs remains a challenge, due to insufficient quality data containing sufficient features. With sufficient training data where all spatiotemporal patterns are well-represented, existing GNN models can make reasonably accurate predictions. However, existing methods fail when the training data are drawn from different circumstances (e.g., traffic patterns on regular days) compared to test data (e.g., traffic patterns after a natural disaster). Such challenges are usually classified under domain generalization. In this work, we show that one way to address this challenge in the context of spatiotemporal prediction is by incorporating domain differential equations into Graph Convolutional Networks (GCNs). We theoretically derive conditions where GCNs incorporating such domain differential equations are robust to mismatched training and testing data compared to baseline domain agnostic models. To support our theory, we propose two domain-differentialequation-informed networks called Reaction-Diffusion Graph Convolutional Network (RDGCN), which incorporates differential equations for traffic speed evolution, and Susceptible-Infectious-Recovered Graph Convolutional Network (SIRGCN), which incorporates a disease propagation model. Both RDGCN and SIRGCN are based on reliable and interpretable domain differential equations that allows the models to generalize to unseen patterns. We experimentally show that RDGCN and SIRGCN are more robust with mismatched testing data than the state-of-the-art deep learning methods.