Home | Publications | TSS+26

Explainable AI for Streaming Process Mining: Overview, Challenges and Outlook

MCML Authors

Gabriel Marques Tavares

Dr.

→ Group Thomas Seidl
Database Systems, Data Mining and AI

Udo Schlegel

→ Group Thomas Seidl
Database Systems, Data Mining and AI

Thomas Seidl

Prof. Dr.

Director

Database Systems, Data Mining and AI

Abstract

Streaming Process Mining (SPM) enables the real-time analysis of business processes through the continuous processing of event streams, providing timely insights and facilitating quick decisions. Although SPM supports various applications in multiple domain areas, it often relies on data-driven algorithms that lack interpretability, limiting user acceptance and trust. To tackle such shortcomings, explainable Artificial Intelligence (XAI) offers promising techniques to improve transparency and understanding of the applied approach. Despite their complementarity, SPM and XAI originate from distinct research communities, each with its own focus, methodologies, and evaluation criteria. This paper provides an overview of the intersection between XAI and SPM. We categorize existing work according to the SPM tasks addressed, the XAI methodologies used, and the level of interpretability from the end-user’s perspective. In addition, we identify key research challenges and propose future directions aimed at developing more human-centered and trustworthy SPM approaches. Our findings emphasize the importance of aligning algorithmic performance and human comprehension to support responsible decision making in dynamic process environments.

inproceedings TSS+26