Sizhe Xu is a Ph.D. student in Computer Science advised by Professor Zhaonan Wang. He received his M.S. in Urban Data Science at CUSP, NYU Tandon, where he conducted mobility research advised by Professor Takahiro Yabe. His research focuses on spatially grounded world models and visual representation learning. In particular, he is interested in incorporating geographic information and spatiotemporal dynamics into world models, enabling them to better understand, predict, and simulate complex real-world environments.
Publications
- Sizhe Xu*, Liang Renzhao*, Chenggang Xie, Jingru Chen, Feiyang Ren, Shu Yang, and Takahiro Yabe. "Abstain Mask Retain Core: Time Series Prediction by Adaptive Masking Loss with Representation Consistency." Advances in Neural Information Processing Systems 38 (2026): 99445-99471.
- Li, Boyang, Yulin Wu, Sizhe Xu, Nuoxian Huang, Zhonghang Yuan, Shangyi Guo, Shu Yang, and Takahiro Yabe. “nD-RoPE: A Generalized RoPE for n-Dimensional Position Embedding.” Proceedings of the 43rd International Conference on Machine Learning, 2026.
- Sizhe Xu*, Li Boyang*, Donghak Lee, Shangyi Guo, Zhaonan Wang, and Takahiro Yabe. “Behaviorally Grounding Large Language Models with Revealed Choices for Human Mobility Prediction.” Proceedings of the 34th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, 2026. https://doi.org/10.1145/3841645.3842972.
Research Interests
- Spatiotemporal AI
- Representation Learning
- World Models