AbstractMine

AbstractMine aims at providing the algorithmic foundations to incorporate privacy guarantees in the abstraction of events that are continuously generated by distributed sources. To this end, we first develop a formalism to describe the distributed abstraction of events and link it to attacks on the privacy of process participants as well as the utility of the resulting data for operational analysis. Based thereon, we contribute algorithms for privacy-aware abstraction of events that satisfy predefined privacy guarantees, while maximizing the data utility. While we devise these algorithms first for a centralized setting, we then lift them to a distributed setting by incorporating the data exchange between the involved entities. Finally, we strive for realizations of our techniques that enable continuous operation over streams of events as well as decentralized execution, where all decisions on event abstraction and data exchange are taken locally by each involved entity. Within SOURCED, we work together with ScalableMine on the definition of streaming operators, as well as on benchmarking of our solutions to continuous abstraction and exchange. Moreover, we will align our models for adversaries and privacy threats with those of PrivateMine, also incorporating their results on distributed evaluation of process mining functionality for the distribution of abstraction operators. Our privatization strategies for noise insertion will further inform the noise perturbation mechanisms in ExplainableMine.

Verantwortlich für die Redaktion: Christian Imenkamp