Research fields
Core research areas 1: Design of Intelligent Information Systems
This core research area investigates the socio-technical design of intelligent information systems using modern data science technologies. For example, algorithmic approaches from the fields of deep learning, generative AI, process mining, natural language processing or computer vision are used to develop adaptive and efficient information systems. The aim is to create innovative solutions that take into account both technical/algorithmic and social/organizational requirements.
Research methods:
- Design science research as an overarching approach
- Architecture analysis and system design
- Development and evaluation of prototypes
- Case studies and simulation-based experiments
- Data-driven modelling and optimization
Core research areas 2: Interpretable Machine Learning & Explainable Artificial Intelligence
This core research area focuses on the development and evaluation of intrinsically interpretable machine learning models ("Interpretable Machine Learning", IML) and post-hoc analytical methods for the subsequent explainability of complex black-box models ("Explainable Artificial Intelligence", XAI). The aim is to increase the traceability and transparency of AI systems in order to ensure their use in critical application areas (e.g. medicine, production, finance). In addition, the core research areas deal with all tangential topics that arise from an algorithmic/technical, social/user-centric and organizational perspective.
Examples of related research topics:
- Linking IML and XAI approaches with generative AI or large language models (e.g. in the form of conversational agents as a personalized dialog interface between model and user)
- Personalization of IML and XAI outputs to individual user preferences, taking into account various constraints (e.g. capabilities, traceability, generalizability, etc.)
- Operationalization of the difficult-to-measure construct "interpretability" from a socio-technical perspective as well as possible linkage potentials to established methods from the field of NeuroIS research
- Conducting user-centered experimental and evaluation studies, in particular taking into account theoretical explanatory approaches from the field of behavioral economics (key point: "Thinking, Fast and Slow")
- Building bridges to the subject area of "Causal Machine Learning" to uncover and consider causal relationships
Research methods:
- Development of new algorithmic methods
- Theoretical analysis of interpretability measures
- Benchmarking of explainability techniques
- User studies for the evaluation of explanations
Ongoing projects
- BMBF junior research group: "White-Box-AI - Transparent decision support through interpretable machine learning models: Development and evaluation of interpretable model structures with the inclusion of expert knowledge" (together with Mathias Kraus, University of Regensburg; duration: 08/2022 - 07/2025)
Core research areas 3: Applied Machine Learning
This focus area investigates the application and benchmarking of modern machine learning algorithms in various domains. By evaluating models in real use cases, insights are gained into their capabilities, scalability and robustness. This enables the development of industry-specific adaptations and optimizations for the practical use of AI.
Research methods:
- Empirical evaluation of ML algorithms
- Benchmarking and performance analysis
- Transfer learning and domain adaptation
- Case studies in specific application areas
Core research areas 4: Human-AI Interaction
This core research area deals with the human perception and use of AI systems. User-centered experimental studies are used to analyze factors such as trust, acceptance and decision-making behavior towards AI. The aim is to derive design principles for human-centered AI systems that enable effective and sustainable interaction between humans and machines.
Research methods:
- Laboratory experiments and online studies
- Psychological and behavioral analyses
- Questionnaires and qualitative interviews
- Perspective: eye-tracking and neurophysiological measurements