Our team focuses on basic research topics in the context of Bilateral Software Engineering which focus on exploiting the synergy effects between Software Engineering and AI techniques, including large language models (LLMs). The research of the Applied Software Engineering & AI research group (ASE) can be summarized as follows:

XAI Approaches in Recommender Systems: Explainable AI (XAI) is based on the idea of intelligent explanation concepts that make recommendations and AI-generated results transparent for different stakeholders involved in a recommendation process. Our research also investigates explanation approaches for (LLM-enhanced) recommender systems. Applications include different domains such as Software Engineering (e.g., release planning), Medicine (e.g., medical treatments), and Financial Services (e.g., recommendation of loans).

Recommender Systems: Development of machine learning and LLM-enhanced techniques that help to predict and elicit preferences of users in diverse application scenarios such as Financial Services, Intelligent Video Segmentation, E-learning, and Software Engineering (e.g., Requirements Engineering). This includes the integration of LLMs with traditional recommender-system approaches, for example, for preference elicitation, knowledge acquisition, recommendation generation, and explanation. A specific research focus of our team is to analyse the potential of recommender systems in achieving the 17 Sustainable Development Goals (SDGs) defined by the United Nations.

Configuration Systems: Logic-based and machine-learning-based AI techniques for efficient solution search in constraint-based reasoning and beyond (e.g., model-based diagnosis and conflict detection). A particular focus is on combining LLMs with symbolic reasoning approaches, for example, to support knowledge acquisition, constraint generation, configuration, diagnosis, and explanation. Applications include different domains such as automotive, Software Engineering (e.g., release planning and open-source development), and the Internet of Things (IoT).

Requirements Engineering: Development of machine learning approaches to improve the quality of Software Engineering processes. Examples thereof are automated dependency detection in Requirements Engineering, intelligent Release Planning, and Requirements Recommendation (e.g., for reuse purposes). A specific focus is the exploitation of large language models (LLMs) to boost the quality of Requirements Engineering processes. This includes, for example, requirements elicitation, analysis, recommendation, consistency checking, and knowledge extraction, as well as aspects such as gender awareness and fairness.

Psychology-Aware Artificial Intelligence: A major focus of our research group is the analysis and integration of theories from cognitive psychology into the development of software systems. Examples thereof are persuasive technologies for software development environments, decision-bias-aware recommendation technologies, and psychologically aware recommendation technologies for e-Learning scenarios. In this context, we also investigate how LLM-based and conversational AI systems can take psychological factors into account and support users in making better-informed decisions.

AI methods for Intelligent Learning Support: In this context, we focus on the development of intelligent AI technologies for counteracting forgetting, i.e., machine-learning-based approaches that help to tackle the challenges imposed by forgetting curves. This also includes the exploitation of LLMs for personalized learning support, content generation and explanation.

Overall, generative AI (specifically, LLMs) are not regarded as a separate research topic but as an additional enabling technology that cuts across recommender systems, configuration, requirements engineering, psychology-aware AI, XAI, and learning support.