(from left) Mr LIANG Shen, Prof. XU Yang, Mr LI Guangyue and Mr LUO Yuxiao

A team of four PhD students from the Department of Land Surveying and Geo-Informatics at PolyU brought home the championship at the 14th ACM SIGSPATIAL Cup (GISCUP) in 2025, an annual contest in geospatial AI.

The winning team, named “Mobility Science Lab (POLYU)”, comprised Mr LIANG Shen, Mr LI Guangyue, Mr LUO Yuxiao and Ms WANG Jinghan. It was led by Prof. XU Yang, Associate Head (Partnership) and Associate Professor of PolyU’s Department of Land Surveying and Geo-Informatics. 

The 2025 competition focused on “Prediction of Human Mobility Patterns in Multiple Cities”. This challenge required participants to forecast complex mobility trajectories across diverse metropolitan areas. To tackle the challenge of adapting to various urban layouts, the team developed the “HuMob Predictor” framework. Their innovative approach combined a multi-level encoding module, designed to capture intricate urban structures, with an incremental training strategy that progressively learns from simple patterns to complex dependencies.

The team’s solution demonstrated outstanding stability and performance during the official evaluation, setting a new benchmark for individual mobility prediction. This remarkable achievement underscores the Department of Land Surveying and Geo-Informatics at PolyU’s expertise in addressing complex spatiotemporal problems and advancing generalisable geospatial AI.

GISCUP 2025 was held as part of the ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems 2025, the 33rd edition, held in Minneapolis, United States, from 3 to 6 November 2025. The result of the contest was announced during the conference.