Privacy-preserving Process Mining

Speaker: Stephan Fahrenkrog-Petersen
         Humboldt University of Berlin
         Germany

Title:   "Privacy-preserving Process Mining"

Date:    Monday, 20 December 2021

Time:    4:00pm - 5:00pm

Venue:   Lecture Theater F (near lift 25/26), HKUST

Abstract:

Process Mining is an emerging subfield of data mining, focusing on the
data-driven analysis of business processes. It uses event data recorded
while executing the business process. Here, each execution of an activity
of the process is captured by an event. A sequence of events, referred to
as a trace, then captures the behaviour of single process instance. Yet,
traces may enable conclusions on sensitive information of individuals,
such as patients, customers, or process workers. Anonymizing traces is
challenging, though, since behavioural characteristics need to be
preserved for process analysis. Recently, several techniques have been
developed to address this challenge. In this talk, we review the state of
the art to anonymize event data in Process Mining that ensures
differential privacy. Specifically, we present SaCoFa, a technique that
anonymizes the control-flow of a business process, while considering the
semantics of the injected noise. In addition, we introduce PRIPEL, an
approach that enables publishing of multi-variate event data by exploiting
local differential privacy.


*******************
Biography:

Stephan Fahrenkrog-Petersen is a final-year PhD student, working on
Privacy-preserving Process Mining, an emerging subfield of Data Mining, at
Humboldt-Universität zu Berlin, Germany. His research was published in the
proceedings of the premier conferences in the field (BPM, CAiSE, ICPM) and
in international journals, such as ACM TMIS and KAIS. His work was
recognized with the Distinguished Paper Award at CAiSE 2021 and the Best
Student Paper Award at ICPM 2021.