Combining Differential Privacy and Mutual Information for Analyzing Leakages in Workflows

Autor: Martin Pettai, Peeter Laud
Rok vydání: 2017
Předmět:
Zdroj: Lecture Notes in Computer Science ISBN: 9783662544549
POST
DOI: 10.1007/978-3-662-54455-6_14
Popis: Workflows are a notation for business processes, focusing on tasks and data flows between them. We have designed and implemented a method for analyzing leakages in workflows by combining differential privacy and mutual information. The input of the method is a description of leakages for each workflow component, using either differential-privacy- or mutual-information-based quantification whichever is known for the component. The differential-privacy-based bounds are combined using the triangle inequality and are then converted to mutual-information-based bounds. Then the bounds for the components are combined using a maximum-flow algorithm. The output of the method is a mutual-information-based quantification of leakages of the whole workflow.
Databáze: OpenAIRE