AUTHOR=Sun Yue , Chen Chao , Xu Yuesheng , Xie Sihong , Blum Rick S. , Venkitasubramaniam Parv TITLE=On the generalization discrepancy of spatiotemporal dynamics-informed graph convolutional networks JOURNAL=Frontiers in Mechanical Engineering VOLUME=10 YEAR=2024 URL=https://www.frontiersin.org/journals/mechanical-engineering/articles/10.3389/fmech.2024.1397131 DOI=10.3389/fmech.2024.1397131 ISSN=2297-3079 ABSTRACT=

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 that contains 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) than 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-differential-equation-informed networks: Reaction-Diffusion Graph Convolutional Network (RDGCN), which incorporates differential equations for traffic speed evolution, and the 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 allow the models to generalize to unseen patterns. We experimentally show that RDGCN and SIRGCN are more robust with mismatched testing data than state-of-the-art deep learning methods.