Semantics-Aware Process Mining: Benefits, Current State, and Opportunities
Speaker:
Han van der Aa
Data dell'evento:
Tuesday, 12 May, 2026 - 14:30
Luogo:
Aula Magna, DIAG
Contatto:
Andrea Marrella (marrella@diag.uniroma1.it)
Abstract: Process mining leverages event data to analyze and improve the execution of business processes, supporting tasks such as process discovery, anomaly detection, and next-activity prediction. Traditionally, these techniques have relied on frequency-based methods, focusing on how often certain activity sequences occur in event logs. A growing body of research now explores semantics-aware process mining, which incorporates the meaning of activities — often expressed through textual labels — into analysis, driven by advances in large language models (LLMs). In this talk, Prof. Van der Aa will discuss the potential benefits of semantics-aware process mining, present results from current experiments on few-shot learning, fine-tuning, and instruction tuning, and outline ongoing research directions and further opportunities in this area.
Bio: Han van der Aa is a Professor of Process-Oriented Information Systems at the Faculty of Computer Science of the University of Vienna. His research centers on process mining—the data-driven analysis of organizational processes—with the aim of understanding, analyzing, and improving how organizations operate. In this context, he is particularly interested in considerations of uncertainty, data privacy, applications of natural language processing, and the provision of results that are meaningful to users.
Before joining the University of Vienna in 2024, Han van der Aa obtained a PhD from the Vrije Universiteit Amsterdam in 2018 and previously held positions at the Humboldt-Universität zu Berlin and the University of Mannheim. His work has resulted in close to 100 peer-reviewed publications, which have won awards at flagship conferences in the field, such as CAISE, BPM, and ICPM.
gruppo di ricerca: