ASE Research Group
Better Software. Smarter AI.
How can AI transform the way we build software – and how can Software Engineering make AI more reliable? The TU Graz Applied Software Engineering & AI Research Group (ASE) explores both directions. We combine Large Language Models, AI Agents, Machine Learning, and Constraint Reasoning to address challenges in software and AI systems development.
This two-way exchange between Software Engineering and Artificial Intelligence is at the heart of Bilateral Software Engineering. Bilateral Software Engineering investigates the reciprocal integration of AI and Software Engineering: using AI to advance Software Engineering methods and processes, and using Software Engineering principles to develop, test, and maintain reliable AI systems.
Explore our publications · Discover our projects · Meet our team
Our Research
AI for Software Engineering (AI4SE). We explore how Machine Learning and Generative AI can improve Software Engineering, from requirements elicitation and quality assurance to Vibe Programming, automated UI/UX analysis, and LLM-based recommendations for improving software artefacts (example).
Development of Human-Aware AI Systems. We combine theories of human decision-making with recommendation algorithms for complex domains such as software engineering, medicine, and financial services. Our research focuses on explainable algorithms for real-world applications (example).
Software Engineering for AI (SE4AI). AI systems require clear requirements and effective quality assurance. We investigate how software engineering methods can support the development, testing, and continuous improvement of AI-based systems and agent-driven software development (example).
From Research to Practice
Automated Usability Analysis with AI. We develop AI-assisted methods that analyse software user interface artefacts to identify usability issues and recommend actionable improvements. Our solutions help companies to reduce the time and effort required for usability analysis while making professional usability evaluation accessible to SMEs with limited resources.
Automated Quality Assurance in Requirements Engineering. The rise of generative AI increases the need for clear, consistent, and complete requirements, including those specified through LLM prompts. We develop techniques to automatically improve requirements quality. This helps teams to early identify ambiguities, inconsistencies, and potential defects.
Automated Testing and Debugging of AI Systems. Testing and debugging AI-based systems requires substantial effort. We develop methods to automatically generate test cases, detect and analyse unintended behaviour, and support targeted repairs. Knowledge-based Guardrails help to effectively control unintended AI behaviour in real-world application settings.
AI-based Learning and Knowledge Sharing – An Industrial Application. KnowledgeCheckR is an ASE-developed system that combines recommendation technologies and gamification to support personalised learning and knowledge sharing (with applications in medicine and public administration).
Connect with ASE
Research advances through collaboration. Whether you are exploring a scientific question, addressing an industry challenge, or looking for a student thesis topic, we welcome the opportunity to exchange ideas.
Meet our team · Discuss a collaboration· Explore thesis opportunities