He Digitalizes Food Processing  [27.01.22]

Jun.-Prof. Dr. Christian Krupitzer | Photo: Uni Hohenheim / Corinna Schmid | more press photos in print quality

Since 1 October 2020, he has headed the Department of Food Informatics at the University of Hohenheim - at the interface of Food Science, Biotechnology, and Informatics. He wants to further automate food processing and thus improve not only traceability, but also hygiene and efficiency. In teaching, he relies on blended learning concepts.

Mr. Krupitzer, your department is called Food Informatics. What does that mean?

So far, the term Food Informatics has not been statically defined in research. At the University of Hohenheim, we understand it to mean the digitalization of food processing as well as the food supply chain. With Food Informatics, we operate interdisciplinarily at the interface of Food Science, Biotechnology, and Informatics.

We push the digitalization of food processing in a holistic approach that includes the food supply chain, retailers/distributors and consumers. In close cooperation with partners from industry, we apply methods and technologies from the fields of artificial intelligence, predictive data analysis, Industry 4.0, adaptive software systems, and the Internet of Things (IoT) to food processing. The goal is to increase sustainability, but also efficiency.

Could you please tell us about your own career?

After my Bachelor's and Master's studies in Information Systems at the University of Mannheim (with an Erasmus semester in Amsterdam), I did my PhD at the Department of Information Systems under Prof. Dr. Christian Becker in Mannheim. My dissertation topic was in the area of software engineering for adaptive software systems. Self-adaptive software systems are systems that can adaptively adjust their behavior during runtime to respond to changes in their environment, such as self-driving vehicles.

After my doctorate in August 2018, I worked as a group leader for the research group "Internet of Things and Cyber-physical Systems" with Prof. Dr.-Ing. Samuel Kounev at the University of Würzburg until I started my position at the University of Hohenheim in October 2020.

What is the most important question that drives you in your research?

Food processing has long been highly automated, particularly for reasons of hygiene - the fewer people there are on the industrial floor, the less likely the products are to be contaminated with bacteria introduced by humans. But even here, there is still room for increased efficiency, especially by means of digitalization and consequently better control of the processes through data analysis. Adaptive control of the processing process is one of the major research topics in our department.

Traceability is another important topic. The goal is to have a complete overview from the origin of the raw materials to the supermarket ("From Farm to Fork"). This is not only in the interest of companies for reasons of maintaining product quality. Customers are also increasingly demanding transparency. Food Informatics has a big part to play in making this possible.

If you had unlimited resources and possibilities at your disposal: Which project would you tackle?

End-to-end traceability "From Farm to Fork" combined with machine learning to proactively detect critical conditions would be an exciting project. This is currently only possible to a limited extent, as the various players do not exchange data and many more sensors would be needed to get a complete picture of the supply chain.

Data exchange and the use of many different sensors and analyses, which are often dispensed with today for cost reasons, would facilitate the creation of a digital image of the food, the so-called digital twin, which could enable the simulation and optimization of food processing.

What research topics are you working on at the moment?

Currently, we are researching two topics in particular: Food Supply Chain and Digital Twin.

In a current project, we are focusing on the intelligent integration of diverse technologies - such as smart sensors, machine learning, blockchain - within the food supply chain to support food traceability and safety. Novel sensors integrated into the packaging design contribute to the targeted collection and storage of relevant data in the blockchain.

Real-time analytics are enabled by integrating current machine learning trends "on the edge" (Edge ML), meaning that the analytics take place directly at the point of origin of the data, rather than retrospectively in the cloud. This also allows forecasting to be integrated, for example, to anticipate critical conditions predictively and act adaptively.

With the help of a digital food twin, the current status within production can be tracked at any time. While approaches from Industry 4.0 often focus on the analysis of machine data, we also aim at product-related data analysis (e.g. the effects of pressure exerted by machines).

By means of 3D printing, replicas of food products are to be generated which - when equipped with sensors - allow conclusions to be drawn about the processing steps involved. Machine learning techniques will be used to generate the digital twin from this data and other data sources (e.g. machine data, process data or raw material data), thus ensuring traceability of current production and food status, but also enabling simulation of the variability of the food in the processing process.

Can students participate in your research projects?

I generally try to practice research-based teaching. This starts with integrating current research work in our department into the lectures as case studies. But seminar papers, Humboldt Reloaded or Master's projects, and Bachelor's/Master's theses are also seamlessly integrated into our research.

In a cooperative supervision approach, students are shown from the beginning how their work will be integrated into our research. High-quality papers are published jointly with the students, i.e. they can actively contribute to the publications as co-authors. Likewise, students in HiWi jobs support our research by conducting research or contributing to the implementation of software prototypes.

What are the main contents of your teaching concept?

In our courses we rely on blended learning concepts. This means that the knowledge traditionally imparted in lectures through frontal teaching is conveyed in videos. This is supported by self-learning questions. In face-to-face events, we get into the topic with the students and conclude the teaching units with a quiz, a case study, or exercises, for example. This is supported by very practical exercise units.

The decision for the blended learning concept was made before Covid, because with such a concept students can learn at their own pace, e.g. by pausing the video and getting more information via other websites. Also, such videos support the teaching of programming skills, as students can write the code in parallel and "experiment" with variations.

What does good teaching mean to you personally?

To me, the applied transfer of knowledge is particularly important. On the one hand, this means thinking carefully about what the goal of the course should be and designing the course accordingly. For example, I'm not a fan of pure knowledge testing in exams, but rather rely heavily on transfer tasks.

On the other hand, it must also be clear what the target group is and the content must be adapted accordingly. One example is our course "Fundamentals of Computer Science" in the Bachelor's program: We are aware that we will not shape the students into software developers and that this is not the students' goal. However, both we and the students see the need to acquire programming skills in order to be able to use them later for data analysis in the context of the final thesis or in professional life. That's why we focus on this aspect and specifically teach the skills that are necessary for this.

Where do your graduates work later?

The career prospects are very good: 84 percent of companies in the German food industry see digitalization as a great opportunity. This is one of the findings of the 2019 survey "Die Ernährung 4.0 - Status Quo, Chancen und Herausforderungen" (Nutrition 4.0 - Status Quo, Opportunities and Challenges) conducted by the digital association Bitkom and the Federation of German Food and Drink Industries (BVE). In particular, the current decade is seeing an increased focus on scenarios that require the integration of digitalization, e.g., order production in real time through digitalized process control (62% of the companies in the survey) or production of individualized foods through intelligent process control (65%).

However, the aforementioned survey by BVE and Bitkom also shows that a lack of digital competence among employees (88%) is a major obstacle. Employees must combine knowledge of both machine learning and the domain, i.e., the processes/products, in particular to support automated data analysis. This is necessary because analytics using machine learning only identifies correlations, but expertise is required to interpret causality.

Such personnel are hard to find, as the typical (even university) education in Food Technology does not yet sufficiently integrate Information Technology. For this reason, data scientists with specialized knowledge - in this case in food processing - are currently in high demand on the job market.

What good advice would you like to give to students?

The most important skill as a Food Informatics graduate is certainly interdisciplinarity. In their day-to-day work, they have a mediating function between the specialist departments that manage the food processing processes and the IT department as the technical point of contact. For this purpose, it is important that Food Informatics specialists have knowledge in both areas.

Analytical thinking and structured work are also important. Both sub-disciplines, Food Technology and Computer Science, need these skills per se: Food Technology thrives on structured procedures, e.g. in laboratory experiments, and in Computer Science analytical problem solving is particularly relevant. In Food Informatics, these are of outstanding importance, as Food Informatics specialists consider one of the most important goals to be the optimization of processes.

How do you like it here in Hohenheim?

Unfortunately, the Covid pandemic made it very difficult to get started. But in particular, the very structured preparation of information on the intranet and Mr. Klumpp as a contact person for new recruits supported me very well. The helpfulness of the employees in the administration and their availability for discussions even during times when everyone was working from home was also very great.

I was also very well received among the professors. I particularly like the fact that I have never had the feeling that there is a difference between junior professors and "normal" professors. The positive feedback from students after the first courses was of course also a nice confirmation. And in the end, despite working from home, our small team has now settled down very well and we have a positive, creative atmosphere in the department.

One last question, Mr. Krupitzer: How do you spend your free time?

I used to go jogging regularly and also participated in 10-km runs, half marathons and marathons - with the highlight being participation in the Berlin Marathon. But as a father of three young children aged one, three, and five, my free time is now mostly taken up by the little ones ;)

Thank you very much for the interview, Mr. Krupitzer!

Back to Topic service