[ PMUD ]

Workshop overview

The Process Mining with Unstructured Data (PMUD) workshop aims to provide a forum for researchers and practitioners to present and discuss how unstructured data can support process mining tasks.

Traditional process mining techniques take structured data as input. However, many valuable insights can be hidden in unstructured data sources, such as emails, social media interactions, legal documents, images or sensor data. Most state-of-the-art techniques ignore such data, thus missing valuable insights regarding the process. Furthermore, relying solely on structured data can lead to a rigid analysis framework, as structured data often adheres to predefined formats and categories.

Recently, a growing array of approaches to deal with various kinds of unstructured data has emerged in the literature. Among them, NLP techniques have attracted considerable interest thanks to recent breakthroughs such as Large Language Models. Examples include using NLP techniques to extract process models from textual documents, or using LLM to interact with the user at runtime. Some studies also advocate using unstructured data to extract inter-case patterns, thus supporting process-level and object-centric approaches.

Despite the promising results, dealing with unstructured data remains one of the main challenges when applying process mining.

Topics of Interest

We welcome contributions on leveraging unstructured data to develop new or improve existing techniques, frameworks and tools to support the whole spectrum of process mining tasks.

This includes (but not limited to) the following topics:

• Leveraging unstructured data for multi-perspective process discovery

• Leveraging unstructured data for object-centric process mining

• Conformance checking against unstructured constraints

• NLP for conformance diagnostics

• Unstructured data for model enhancement

• Leveraging unstructured information for intra-and inter-case predictive analytics

• NLP for explainable Predictive Process monitoring

• NLP for interactive recommendations

• Comparative process mining on unstructured process data

• LLMs for process mining tasks

• Case studies on the use of unstructured data for process mining tasks