Real-world spatiotemporal systems—such as urban transportation, maritime, and low-altitude aviation economies—generate large, heterogeneous, and continuously evolving data streams with strong spatial dependencies, long-range temporal dynamics, and significant operational constraints. While recent advances in AI-based modeling have shown promising results on benchmark datasets, many approaches remain insufficiently grounded in real operational settings. This half-day workshop focuses on AI-enabled solutions for real-world spatiotemporal systems, with particular emphasis on transportation, maritime networks, and low-altitude aviation economies. The workshop brings together researchers and leading domain experts with deep system-level knowledge to present real operational challenges, data limitations, and decision-making constraints through invited keynote and expert talks. Building on these real-system perspectives, the workshop explores how modern learning methods—such as deep neural models, graph and sequence representations, state-space approaches, and constraint-aware modeling—can be designed to meaningfully resolve practical spatiotemporal problems, supporting robust prediction and decision-making in complex systems.
Workshop Organizers
- Junbin Gao
- Michael G. H. Bell Zhiqi Shao
- Zhiqi Shao
- Haoning Xi
- Ze Wang
- Dr. Shoujin Wang
- Jiayu Fang