Faizan Ahmed Khan - From Event-Log Abstraction to Routine Discovery in Robotic Process Automation

This presentation explores how to extract meaningful process models from low-level data. I first introduce a clustering-based approach for event-log abstraction, which improves the interpretability of process models. I then extend this idea to the domain of Robotic Process Automation (RPA), where routines must be discovered from unstructured UI logs.

Unlike traditional event logs, UI logs present additional challenges, including the absence of case identifiers, interleaving of routines, and shared actions across multiple routines. To address these issues, I present approaches based on segmentation, encoding, clustering, and graph-based techniques.

The results demonstrate how abstraction-driven methods can effectively bridge low-level data and high-level process understanding in real-world RPA settings.

Faizan Ahmed Khan is a final-year PhD candidate in Brain Mind and Computer Science at the University of Padova, Italy. His research focuses on process mining, with particular emphasis on event-log abstraction and routine discovery in Robotic Process Automation (RPA). His work investigates how to extract meaningful process knowledge from low-level and complex data sources such as UI logs.