Temporal Spatial Causal

Temporal-spatial causal inference aims to understand cause-and-effect relationships across both time and space, addressing challenges like spatial interference where a treatment's impact extends beyond its immediate location. Current research focuses on developing deep learning models, often incorporating graph neural networks and latent factor modeling, to estimate direct and indirect causal effects from spatiotemporal data, improving interpretability through techniques like causal graph generation. This field is crucial for analyzing complex systems in domains like climate science and traffic prediction, enabling more accurate forecasting and a deeper understanding of dynamic processes.