Zhongfu Ma Wins Second Place in the GISS-SG Student Honors Paper Competition at AAG 2025

Zhongfu Ma, our PhD student, received second place in the 2025 Geographic Information Science & Systems Specialty Group (GISS-SG) Student Honors Paper Competition, awarded at the American Association of Geographers (AAG) Annual Meeting in Detroit. The GISS-SG competition recognizes outstanding student research on theoretical, conceptual, and methodological developments in GIS as well as on particular innovative GIS applications, attracting top student researchers nationwide. This annual event is treated as one of the most competitive student paper competitions in AAG, and is followed by The Waldo Tobler Distinguished Lecture in GIScience that attracts hundreds of scholars in the field of GIS and Spatial Data Science.

The awarded paper, "Gravity-Informed Deep Flow Inference for Spatial Evolution in Panel Data", proposes a novel deep learning model, Deep Spatial Evolution Network (DSEN), to infer spatial evolution, i.e., flows, underlying dynamic spatial distribution. The model integrates geographic context across events using Graph Attention Networks and incorporates a gravity-informed spatial evolution decoder for flow inference. Dr. Di Zhu is the corresponding author of this paper.

This research is validated on large-scale individual mobile positioning data from the Twin Cities Metropolitan Area to infer human mobility flows during the Christmas holiday in 2021. The model achieves superior performance among the baselines, and the evolutionary features extracted from the model provide novel insights for understanding human mobility from a panel perspective.