Current statistical methodology still reflects the use-context of its first development in the 1920s: methods assume that researchers make all important analysis choices before gathering the data. In modern data science age, however, data-driven paradigms have become dominant: many measurements are gathered simultaneously and it is often natural to pursue research questions that were not of obvious interest a priori. Looking at the data is scientifically sensible to do, but the problem is that current statistical methods cannot handle this behaviour well. What is needed for this, is the right mathematical framework, that gives the necessary error control guarantees without all the old restrictive assumptions. In this PhD project you will develop such mathematical foundations by bringing together exciting novel developments in mathematical statistics and machine learning theory, among which: closed testing, bandits, and the new theory of hypothesis testing with e-values, co-developed by your supervisor Dr. Rianne de Heide.
You will:
- Perform daily PhD-level research
- Publish results in journals and conference proceedings, and present these at (inter)national workshops and conferences.
- Contribute to teaching activities related to your work.
- Be a part of an excellent young research group, and benefit from frequent (virtual) interaction with other internationally renowned research groups working on e-values, such as at CWI (Amsterdam), LUMC (Leiden) and in the US.
We are an inclusive group and diversity is at the heart of our research principles. We care about a good working atmosphere and a good work-life balance. Applications from all groups currently under-represented in academic posts are especially encouraged. We particularly encourage women and people from minority groups to apply.