PhD School Business Informatics 2023

The PhD School on Business Informatics is a joint project initiated by Prof. Hans-Georg Fill and Prof. Edy Portmann of the Department of Informatics and funded by the rectorate of the University of Fribourg. The goal of the PhD school is to give an introduction into scientific research, the methods used in business informatics and computer science, insights into concrete research activities, and for PhD students to receive feedback on their research projects. This is complemented with talks from experts in neigbouring fields and from industry.

The PhD School is open for all PhD students in business informatics and computer science of University of Fribourg as well as external PhD students upon registration. External PhD students need to submit a recommendation letter from their supervisor upon registration.

Date: September 11-13, 2023

Location: University of Fribourg, Boulevard de Pérolles 90, PER 21 B130, CH-1700 Fribourg

Registration: Online Form

Deadline for Registration: August 1, 2023

Fees: 300 CHF, students of UniFR sponsored by rectorate

Number of participants: Max. 20 (preference given to UniFR PhD students)

 

On the first day,  introductions into scientific research, research methods, paper writing and the management of open science and open data as well as ethics in research are given.

Day 1 Topic Lecturer(s)
9:00 - 9:15 Welcome Hans-Georg Fill, Edy Portmann
9:15 - 10:45 Method Engineering Jolita Ralyté, University of Geneva, CH
10:45 - 11:00 Coffee break  
11:00 - 12:30 Computational Literature Analysis Felix Härer & Hans-Georg Fill, University of Fribourg, CH
12:30 - 14:00 Lunch  
14:00 - 15:30 What Makes (Good) Research? A Software Engineering Perspective Timo Kehrer, University of Berne, CH
15:30 - 15:45 Coffee break  
16:00 - 17:30 Understanding the digital world: Modeling with Heraklit Peter Fettke, German Research Center for Artificial Intelligence and Saarland University, Germany

 

On the second day, further current research activities and methods are presented. At the end of the day, a joint dinner will be organized to which all participating PhD students and professors will be invited.

Day 2 Topic Lecturer(s)
9:00 - 10:30 Paving the way from Interpretability to Trustworthiness in Artificial Intelligence José María Alonso Moral, Universidade de Santiago de Compostela, Spain
10:30 - 10:45 Coffee break  
11:00 - 12:30 Formalizing and interpreting human requirements in evaluation Miroslav Hudec, Technical University of Ostrava, Czech Republic
13:00 - 14:00 Lunch  
14:00 - 15:30 Computer Science: Quo vadis? Kilian Stoffel, University of Neuchâtel, CH
15:30 - 15:45 Coffee break  
16:00 - 17:30 Perceptual Computing  Edy Portmann, University of Fribourg, CH
18:00 - 21:00 Joint Dinner at Brasserie du Commerce  

 

On the third day, the participating PhD students will briefly present the current state of their PhD and receive feedback from the participating professors. Finally, there is also an industry talk showing the perspectives of a PhD in industry.

Day 3 Topic Lecturer(s)
9:00 - 12:00 PhD Student Presentations and Feedback Session (20 min./person: 12 minutes presentation, 8 minutes discussion) Simon Curty; David Huser; Jose Mancera; Erika Minarikova; Fabian Muff; Benedikt Reitemeyer; Miloš Švaňa; Gozde Ayse Tataroglu Ozbulak; Iva Vasic
12:00 - 14:00 Lunch  
14:00 - 15:30 Industry Talk Jhonny Pincay-Nieves, Senior Data Scientist & Developer, Viasuisse AG / SA
15:30 - 16:00 Coffee break  
16:00-17:00 PhD Student Presentations and Feedback Session (20 min./person: 12 minutes presentation, 8 minutes discussion) Alexander Völz; Marcel Bühlmann; Narek Andreasyan
17:00 - 17:30 Final Session and Recap All

 

Abstracts of the Talks:

 

Prof. Dr. Jolita Ralyté, University of Geneva, Switzerland
Method Engineering
The field of information services and systems is constantly evolving thanks to new information technologies such as AI, IoT, Big Data, Digital Twins, Blockchain, etc., which foster innovation and enable new business models. Naturally, they inspire researchers and practitioners to build new methods, frameworks, modeling techniques, design approaches and tools aiming at supporting these new situations. The discipline of Method Engineering (ME) provides principles and systematic approaches for creating, designing, and customizing methods and frameworks for a specific purpose, domain, or organization. The risk of neglecting the theory of ME is to end up with method artifacts that are not smart enough: that do not adapt easily to new situations, lack interoperability or scalability, are too complex and difficult to apply in practice, etc. The lecture will present an overview of ME principles and techniques and provide recommendations on how to build smart methods and how to do it smartly.

 

Prof. Dr. Timo Kehrer, University of Berne, Switzerland
What Makes (Good) Research? A Software Engineering Perspective
Research is an organized process that aims at acquiring new knowledge or insights addressing gaps in our understanding of a specific area of interest. While this high-level description seems obvious, things get more complicated when having a closer look. Research can take various forms depending on the nature of the subject, and different fields substantially differ in their research cultures. For example, have you
ever thought about that natural scientists might wonder about (us) computer scientists in that we are researching a subject that we have created ourselves? So, what is research in computer science and what makes good computer science research? While the talk does not aim to give a comprehensive answer to this question, it aims at reflecting on what makes (good) software engineering research. Or, more precisely, it
reflects on what the community believes is good software engineering research, paired with some personal opinions and experience. To put it bluntly but particularly interesting for young researchers; how do I get my next paper accepted? While the talk is mostly centered around software engineering as a research discipline, there is certainly potential for generalization, (hopefully) yielding some take away messages for all participants of the PhD school.

 

Prof. Dr. Peter Fettke, German Research Center for Artificial Intelligence and Saarland University, Germany
Understanding the digital world: Modeling with Heraklit
Actual hot topics and trends such as big data, cyber-physical systems, and digital ecosystems pose new challenges to proper handling, understanding, manipulating, etc. the digital world. Established modeling frameworks are no longer up to those requirements. A comprehensive modeling framework is required that integrates three decisive aspects: (1) architecture models with components and their composition, (2) a representation of real and imagined world items and data, and (3) the description of dynamic behavior. We suggest Heraklit, a framework with substantially more far-reaching concepts, methods and best practices. The participants will learn how to use Heraklit. Various case studies demonstrate the central concepts and benefits.

 

Prof. Dr. José María Alonso Moral, Universidade de Santiago de Compostela, Spain
Paving the way from Interpretability to Trustworthiness in Artificial Intelligence
Interpretability is a prerequisite for explainability in the context of Explainable AI. Thus, we can scrutinize intelligent systems and verify if they make decisions in agreement with accepted rules and principles. We will discuss 1) how to attain Trustworthy AI as a safe way of translating knowledge into products and services for economic and social benefit; and 2) how to endow human-centric intelligent systems with trustworthiness, being aware of technical but also Ethical and Legal issues.

 

doc. Dr. Ing. Miroslav Hudec, Technical University of Ostrava, Czech Republic
Formalizing and interpreting human requirements in evaluation
People can express their requirements linguistically. It is a vague, but powerful way. Flexibility in formalization allows us to cover various aspects of reasoning, learn parameters from data, and explain the solution. In the lecture, we discuss: 1) elementary requirements, when even a binary one can be linguistically relaxed; and 2) aggregating elementary requirements. In this context, evaluation and classification intertwine each other. It is suitable when we have a smaller data set, privacy should be met, and we evaluate less-frequently examined entities.