Wei Fan
Ph.D.
Postdoctoral Researcher in Machine Learning
Machine Learning, Time Series Modelling, Data-Centric AI, AI for Health Informatics
Academic History
Wei Fan is currently working as a Postdoctoral Researcher in the Medical Sciences Division at the University of Oxford, UK. His research focuses on data-centric AI, time series modelling, and spatial-temporal data mining. He is also dedicated to applying these methods to solve real-world data science applications, such as healthcare, transportation, and energy. Wei earned his Ph.D. degree in Computer Science from the University of Central Florida. He has interned at industry organizations such as Microsoft Research, Baidu Research, and Bytedance AI Lab.
Wei has published over 30 papers in leading machine learning and data mining journals (e.g., TKDE, TKDD) and conferences (e.g., ICLR, NeurIPS, ICML, KDD, AAAI, IJCAI). Two of his papers were selected as spotlight papers of ICLR. Some recent representative papers can be categorised as follows:
Deep Time Series Modelling
- The Data Learning Perspective: ICLR 22, NeurIPS 23, NeurIPS 23, NeurIPS 24
- The Data Manipulating Perspective: AAAI 23, IJCAI 24, IJCAI 24, KDD 25
Spatial-temporal Data Mining and Interdisciplinary Applications
In addition, Wei has also co-organized Workshops at conferences such as ICDM and CIKM. Wei also serves as a PC member/reviewer for conferences and journals such as ICLR, NeurIPS, ICML, KDD, WWW, IJCAI, AAAI, MM, SDM, LOG, AISTATS, PRICAI, IEEE Bigdata, IEEE TKDE, IEEE TCSS, IEEE IoT, IEEE TBD, ACM TKDD, ACM TOIS, ACM TOMM, Scientific Reports.
Wei has served as the Supervisor for the Ellison Institute of Technology Centre for Doctoral Training in the Fundamentals of AI. He would very much like to have research discussions with DPhil students who are interested in the deep learning-based time series modelling and healthcare applications.
Recent publications
-
DEWP: Deep Expansion Learning for Wind Power Forecasting
Journal article
Fan W. et al, (2024), ACM Transactions on Knowledge Discovery from Data, 18, 1 - 21
-
Dish-TS: A General Paradigm for Alleviating Distribution Shift in Time Series Forecasting
Journal article
Fan W. et al, (2023), Proceedings of the AAAI Conference on Artificial Intelligence, 37, 7522 - 7529