Cross-Silo Process Mining with Federated Learning
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Process analytics techniques such as process discovery play an important role in mining event data and providing organizations with insights about the behaviour of their deployed processes. In many practical settings, process log data is often geographically dispersed, may contain information that may be deemed sensitive and may be subject to compliance obligations that prevent this data from being transmitted to sites distinct to the site where the data was generated. Traditional process mining techniques operate by assuming that all relevant available process data is available in a single repository. However, anonymising, giving control access and safely transferring sensitive data across organization/site boundaries while preserving priacy guarantees is non-trivial. In this paper, we lay out the first steps for a federated future for process analytics where organizations routinely collaborate to learn and mine geographically dispersed process-related data.
Open Access Status
This publication is not available as open access