How would you describe your research in one sentence?
My research is concerned with analyzing how algorithms behave on real data and in real usecase scenarios, understanding why they show this behavior, and using that knowledge to build new, better algorithms.
What is your area of research and what sets it aside from other areas in your field?
A large part of my research can best be described by the term “algorithm engineering”: Algorithms undergo an engineering process, they are designed, analyzed, implemented, evaluated, and if necessary, re-designed and so on, until we have the desired outcome. Algorithm engineering bridges the gap between algorithm theory on the one side, where algorithms are designed and analyzed according to theoretical criteria only, and practice on the other side, where an algorithmic problem appears in some real-world application and a working solution is required. I am especially interested in so-called “dynamic algorithms” that can deal with changes to the data they are using, and that can adjust to such changes efficiently. This is particularly important in the context of big data, or everywhere where recomputing everything from scratch is too resource-intense. The resource is most often time, but can also be money if we have to pay per computing hour, or power usage, for example. Because of the massive amount of data involved, having algorithms that scale or help a system to scale is also very important for many AI applications.
Where do you see the connection between your research and everyday life?
Everywhere. Well, almost. I often think about the algorithms that probably work behind the scenes when I do something “normal”, like looking for the best route, booking a table or a hotel, or something else, and I ask myself how these algorithms could potentially be improved.
Since Computer Science is a very broad field, how would you describe it?
It is a very broad field, but that makes it also so interesting and versatile in my view. In many cases, it boils down to the question: What can we “teach” a machine to do (and how)? Where are the limits?