The statistics research group at the University of Twente invites applications for three fully funded PhD positions in statistics. The successful PhD candidates will conduct research in one of the following three research themes:
Statistical theory for deep neural networks – Why is deep learning so powerful and when does it break down? In the past years there has been a lot of progress on building a statistical foundation that can answer these questions. But still many unsolved problems remain.
Statistical methodology for nano-scales – We are working on several challenging statistical problems related to recent advances in nano-optics and cell biology. Statistical methods are developed within interdisciplinary collaborations with the labs. We also develop the underlying statistical theory to prove that these methods indeed perform well.
Statistical learning theory with applications to time series forecasting – Ensemble learning algorithms such as boosting, bagging and random forests are known to enhance computational efficiency while still providing good statistical reconstructions. In this research theme, ensemble learning methods are investigated for time series data with a focus on change-point problems and long-range dependent time series.