"I see exciting points of contact and fields of application".

Prof. Dr. Pascal Kerschke heads new professorship for "Big Data Analytics in Transportation" at the Institute of Transport and Economics

There is a new face among the professors at the "Friedrich List" Faculty of Transport and Traffic Sciences. Prof. Dr. Pascal Kerschke (34) is the head of the new professorship for Big Data Analytics in Transportation at the Institute of Economics and Transport since 15 March 2021.

The faculty members welcome him warmly and wish him a good start.

Shortly before he moved from Münster to Dresden, he was interviewed about his motivation for applying, his research focus and his goals as a newly appointed professor.

Prof. Kerschke, when you read the advertisement for the professorship, you knew straight away: This is my position?
Prof. Kerschke: The orientation of the professorship appealed to me directly, as it fitted my profile very well, especially in terms of methodology. Nevertheless, I first had to look into myself to make the connection between Big Data Analytics and transport and traffic sciences. The more I thought about it, the more attractive and exciting the position became for me.

Where did you see links to your previous research?
Prof. Kerschke: The faculty is increasingly striving towards Big Data and AI in various areas - such as mobility, transport or logistics. The same applies to overarching research projects in the context of autonomous and connected driving or smart cities. Here I see many opportunities to generate real added value together with other professorships at the institute and at the faculty, both methodologically and application-oriented. The solution of current and future transport economic questions is more and more related to my main research areas of data science, machine learning, data analytics or (multi-objective) optimisation.

How does that differ from your previous activities?
Prof. Kerschke: So far, my research has been very methodically oriented. One focus was / is optimisation. Here I am particularly interested in the improvement (increase in efficiency) of the optimisation process itself, as well as the characterisation of the underlying problems, as this results in a better understanding of the problem. I always use different methods from (statistical) data analysis and (automated) machine learning. Unfortunately, it was very difficult to find practice partners for projects, as many companies do not want to make their data available - especially not if they are to be published afterwards as part of a publication.

© Andrea Surma/FVW

You already mentioned your intended collaboration with other researchers in Dresden. To what extent was the local research landscape decisive for your application?
Prof. Kerschke: For me, a lot of positive things come together here: The TU Dresden offers a great environment for scientists, plus the status of an Excellence University, the network around Dresden-concept, the competence centre ScaDS.AI or the Centre for Information Services and High Performance Computing at the TUD, to name just a few. I am very excited to see what will result from this for my research and work.

Besides research, teaching will also be an important area. What would you like to impart to your students?
Prof. Kerschke: I would like to give them a first feeling for data and familiarise them with approaches for working with data. Where do I get data from? What do I do with them when I have them? How can I process them in a meaningful and profitable way? Simple and clear visualisations, supplemented by various methods and procedures from computer science and statistics, are already very helpful. I would also like to try to take away the scepticism and reluctance that often exists on these topics. Experience shows that in many cases the use and interpretation of these methods is much easier than it first appears. Of course, I personally use these methods regularly in my own projects. They help me to recognise possible patterns in the available data, to extract specific information from it and to check the plausibility of my models and results.

So your motto is: Don't be afraid of big data?
Prof. Kerschke: That's right. Through (interactive) exercises and exciting fields of application from practice and later professional fields, I want to pick up the students, create a basis (of knowledge and tools) and encourage them to think actively. I can also imagine alternative teaching formats such as a hackathon at the faculty or a "Big Data Analytics ski camp" - skiing during the day and lectures followed by vivid and exciting discussions in the evening. We had such a format at the University of Münster and it was very well received.

The even stronger international networking is a focus in the faculty strategy adopted in 2020. You have already been abroad as a pupil and student. Where do you get involved internationally as a scientist?
Prof. Kerschke: International networks and conferences are important platforms in my field for exchanging ideas and publishing new research results. I use that wherever possible. For example, I am a member of the advisory board of the international research network COSEAL, which deals specifically with the (automated) selection and configuration of algorithms. I also support the CLAIRE initiative, which aims to promote European excellence in all areas of AI.

You are not only changing your place of work, but also your place of residence. Did you have any previous connections to Dresden and the TUD?
Prof. Kerschke: Few. Several former classmates and a few friends from the past went to Dresden to study. I visited two or three times. But the last time was already a few years ago. I'm looking forward to exploring the city and Saxony beyond the academic world. For my hobbies, cycling and hiking, there are - I've heard - great opportunities here.                                                              

About Pascal Kerschke:
Pascal Kerschke grew up in Frankfurt (Oder). After graduating from high school, including a high school year in California, he first completed a Bachelor's degree in Data Analysis and Management (BSc.) and then a Master's degree in Data Science (MSc.) between 2007 and 2013 - both at the Faculty of Statistics at TU Dortmund University. He spent an Erasmus semester in Bergen (Norway).

This was followed by a doctorate in information Systems (Dr. rer. pol.) in conjunction with a position as a research assistant at the School of Business and Economics at the University of Münster (WWU). He is the winner of the WWU Dissertation Award 2018. Since November 2017, he has led the Machine Learning and Data Science research group at the Chair of Statistics and Optimization there (formerly Information Systems and Statistics). During his time in Münster, he also spent several weeks on research stays with renowned international scientists in Adelaide (Australia), Mexico City (Mexico), Rio de Janeiro (Brazil) and Vancouver (Canada).

His main research areas:

• Data Science
• Exploratory Landscape Analysis
• Algorithm Selection and cCnfiguration
• Black-box optimisation
• Evolutionary and/or multi-objective optimisation
• Machine learning

He is a member of numerous national and international organisations and committees such as the European Research Center for Information Systems (ERCIS), the ACM Special Interest Group on Evolutionary Computation (SigEVO), the IEEE CIS Task Force on Benchmarking, the Society for Classification (GfKl) Data Science Society, the Society for Informatics (GI), the German Statistical Society (DStatG), and CLAIRE (Confederation of Laboratories for Artificial Intelligence Research in Europe). He is also a coordinating founding member of the research networks COSEAL (Configuration and Selection of Algorithms) and Benchmarking Network.