User Tools

Site Tools


pe-bpmn-editor_differentialprivacy

DifferentialPrivacy

This stereotype is reserved for tasks that apply a differential privacy mechanism as part of their computations. It is reserved for future integrations with out differential privacy analyzers and is currently not used for analysis in PE-BPMN.

Required input objects

1…n data objects (data)

Optional input parameters

computation script or (ε, δ) for each input-output pair

Required output objects

1 data object (data)

Restrictions

Restriction: task must have at least one input and exactly one output data objects
Error: at least 1 input and exactly 1 output are required

Restriction: “inputScript” property must exist under stereotype details object - “inputScript” can be empty, but must not have been deleted from model’s file
Error: inputScript is undefined (it can be empty, but cannot be undefined)

Restriction: “delta” property must exist under stereotype details object - “delta” can be empty, but must not have been deleted from model’s file
Error: delta is undefined (it can be empty, but cannot be undefined)

Restriction: “epsilons” property must exist under stereotype details object - “epsilons” can be zeros, but must not have been deleted from model’s file
Error: epsilons are undefined

Restriction: number of inputs or outputs defined (by ID) in the parameter must match with the number of inputs and outputs that are actually present on a model
Error: number of input and output pairs does not match with the number of epsilons saved (some inputs or outputs might have been removed after adding stereotype)

Restriction: epsilon values must not be empty
Error: epsilon values cannot be empty

pe-bpmn-editor_differentialprivacy.txt · Last modified: 2020/03/02 14:47 by pullonen