Research Interests

Robust Statistics (robust inference in non-continuous data).
Robust statistics is concerned with estimation and hypothesis testing when the observations at hand are not identically distributed. A prominent example is the presence of outliers, that is, an unknown proportion of the data follow an unspecified outlier generating distribution. Nearly all concepts from robust statistics assume that the random variables that constitute the observations are continuous (that is, they can attain uncountably many values). I am interested in performing robust statistics when the observations are not continuous, but discrete (i.e. they can only attain at most countably many values).

Mathematical Statistics (empirical processes and high-dimensional models).
Many well-known concepts from theoretical statistics such as the central limit theorem can be described and further generalized by empirical processes. In particular, empirical processes are useful when deriving theoretical guarantees of a certain estimator. In analogous fashion, I use empirical process theory to derive theoretical guarantees (such as concentration inequalities and asymptotic theories) of the estimators I design.
High-dimensional models are models for which the number of variables is (much) larger than the number of observations. This is the case for many modern datasets. Hence, I intend to make sure that the estimators I design also work well in high dimensions.

Computational Statistics and Statistical Computing.
In addition to deriving theoretical guarantees for my methods, I also need to make sure that they are practically applicable by being computable. In particular in high dimensions, statistical estimation tends to be computationally intensive. Hence, there is a need for designing efficient algorithms and fast software implementations for my methods.

Applied Statistics.
Havig an affiliation with a Medical Center, I enjoy applying modern statistical methods as well as the methods I design to datasets from medicine or psychology. I moreover use applied projects for the generation of research ideas: Field-specific statistical problems that one encounters when working with highly-specific datasets (like in medicine) can often be generalized. I identify such problems, generalize them, and solve their generalized versions (theoretically and/or computationally). Thus, an applied project is often a methodological project in disguise.