Guide to Differential Privacy Modifications

1. Front Matter

   Pages i-viii

   PDF

2. Introduction

   • Balázs Pejó, Damien Desfontaines

   Pages 1-4

3. Differential Privacy

   • Balázs Pejó, Damien Desfontaines

   Pages 5-11

4. Quantification of Privacy Loss (Q)

   • Balázs Pejó, Damien Desfontaines

   Pages 13-18

5. Neighborhood Definition (N)

   • Balázs Pejó, Damien Desfontaines

   Pages 19-28

6. Variation of Privacy Loss (V)

   • Balázs Pejó, Damien Desfontaines

   Pages 29-35

7. Background Knowledge (B)

   • Balázs Pejó, Damien Desfontaines

   Pages 37-42

8. Change in Formalism (F)

   • Balázs Pejó, Damien Desfontaines

   Pages 43-49

9. Knowledge Gain Relativization (R)

   • Balázs Pejó, Damien Desfontaines

   Pages 51-54

10. Computational Power (C)

    • Balázs Pejó, Damien Desfontaines

    Pages 55-57

11. Summarizing Table

    • Balázs Pejó, Damien Desfontaines

    Pages 59-77

12. Scope and Related Work

    • Balázs Pejó, Damien Desfontaines

    Pages 79-87

13. Conclusion

    • Balázs Pejó, Damien Desfontaines

    Pages 89-89

Shortly after it was first introduced in 2006, differential privacy became the flagship data privacy definition. Since then, numerous variants and extensions were proposed to adapt it to different scenarios and attacker models. In this work, we propose a systematic taxonomy of these variants and extensions. We list all data privacy definitions based on differential privacy, and partition them into seven categories, depending on which aspect of the original definition is modified.

These categories act like dimensions: Variants from the same category cannot be combined, but variants from different categories can be combined to form new definitions. We also establish a partial ordering of relative strength between these notions by summarizing existing results. Furthermore, we list which of these definitions satisfy some desirable properties, like composition, post-processing, and convexity by either providing a novel proof or collectingexisting ones.

• Data
• Privacy
• Taxonomy
• Differential
• Quantification

• Department of Networked Systems and Services, Budapest University of Technology and Economics, Budapest, Hungary

  Balázs Pejó

• Tumult Labs, Zurich, Switzerland

  Damien Desfontaines

Policies and ethics