Urban Computing and Spatio-temporal Data Mining
Individual Mobility Behavior Modelling
This line modeled how people move through cities, moving from trajectory prediction to map matching, trajectory recovery, privacy-preserving prediction, and generative mobility simulation.
- 2018
- 2019
- 2019
- 2020 PMF: A Privacy-preserving Human Mobility Prediction Framework via Federated Learning #2 Most Cited Paper in UbiComp 2020
- 2020 Learning to Simulate Human Mobility #2 Most Influential Paper in KDD 2020 AI for COVID-19 Track
- 2021
Citywide Spatio-temporal Prediction
This line lifted mobility signals to city-scale dynamics, covering population estimation, crowd flow, traffic flow, network traffic, benchmarks, and universal spatio-temporal prediction.
- 2016 Context-aware Real-time Population Estimation for Metropolis Honorable Mention Award at UbiComp 2016
- 2018
- 2019
- 2019
- 2023 Dynamic Graph Convolutional Recurrent Network for Traffic Prediction: Benchmark and Solution #1 Most Cited Paper in TKDD 2023
- 2024
Operational Urban Decision Systems
This line translates urban computing research into operational decision systems, from traffic signal control work to industry-scale logistics systems at Meituan. It connects spatio-temporal modelling, graph learning, dispatching, order pooling, and assignment optimization with real-world urban operations, including delivery-network work recognized as an INFORMS Edelman Award Finalist.