Differential Privacy (EN)

Concept

Mathematical concept that protects privacy in data analysis

Data flow in Differential Privacy

flowchart LR   A[Original data] --> B[Additive noise]   B --> C[Differentially private result]   C --> D[Publication/Analysis] 

In context

  • Typically used together with Federated Learning
  • Related to: K-Anonymity, L-Diversity, T-Closeness
  • Example use: Google RAPPOR, Apple Differential Privacy
Quelle: AI Generated