Xiaodan ShiAssociate Senior Lecturer

I teach Data Mining and Machine Learning.

My research focuses on deep learning-based mobility prediction and generation, spanning both small-scale/indoor pedestrian trajectories and urban vehicle trajectories.

(1) Trajectory Prediction. Beyond accuracy, my work also emphasizes interpretability and transferability of prediction models, and pedestrian-pedestrian/vehicle social interactions,

(2) Trajectory Generation. Since real-world mobility data often contain sensitive personal information and are sparsely sampled, I aim to create realistic pedestrian and vehicle trajectories at both small and city-wide scales. The generated data approximate the distribution of real trajectories at the aggregation level, supporting research and applications while mitigating privacy concerns,

(3) Energy-related Time Series Forecasting. In addition to mobility, I work on forecasting tasks in the energy domain, including solar energy generation and appliance-level energy usage in buildings.

I have published in top venues such as AAAI, ECCV, ICPR, IEEE Transactions on Intelligent Transportation Systems, IEEE Transactions on Big Data, IEEE Transactions on Neural Networks and Learning Systems, IEEE Transactions on Mobile Computing, Transportation Research Part C: Emerging Technologies, and Applied Energy.

I serve on the program committees of AAAI, CVPR, ICCV, and ACCV, and on the Young Editorial Boards of Nexus and Advances in Applied Energy. My contributions have been recognized with several awards, including the ISPRS Best Young Author Award, the 2021 R&D 100 Awards, and the Smart 50 Awards.