Papers of the Chair of Process Analytics at the BPM Conference

Members of the Chair of Process Analytics, the newest addition to the DWS Group, will have several presentations coming up at the International Conference on Business Process Management (BPM), one of the main conferences in the field, taking place from September 13 to 18. 

Due to this year's virtual set-up, registration for the conference is currently free and open until September 9.

Our members are involved in the following works:

Privacy-preserving event log sanitization (Main track):

Stemming from research conducted at the Humbolt-Universität zu Berlin, this paper presents a new state-of-the-art technique for the sanitization of event data recorded during process execution. The primary goal of the technique is to ensure that published event logs do not disclose personal information about actors involved in a process, while maintaining as much utility for process mining techniques as possible.

IoT-Based Activity Recognition (BPM Forum):

Stemming from research conducted at the DFKI, this paper demonstrates how activity recognition over IoT data can be employed as a means to provide process assistance in emergency response situations where ground robots accompany human team members.

Transforming natural language statements into declarative process constraints

These works both stem from thesis projects conducted by students of the University of Tartu, Estonia. Both provide user-friendly support for the establishment of declarative (i.e., constraint-based) process models. This is achieved through the transformation of natural language statements into declarative constraints. The Demo paper supports this use case through a conversational agent (i.e., a chatbot), whereas the BPM Forum paper allows users to describe models in a speech-based manner. Aside from the application of speech2text, the latter paper also provides considerable improvements in terms of coverage and quality over our previous work.

These papers are both part of the RuM software.