Guide

Guide
1. Introduction
2. Decision tree usage

2.1. Choosing a view
2.2. Interactive tree
2.3. Questionnaire
2.4. Type of problem

3. Scoping
4. Legal considerations

4.1. Legal framework
4.2. GDPR principles in relation to PETs
4.3. Ethics

5. Levels of data-sharing risk mitigation
6. Decision tree guide

6.1. Data sources independence
6.2. Federated Analytics/Learning
6.3. Data vs Model and Output sensitivity
6.4. Set Intersection
6.5. Sensitivity of locally computed values

7. Appendix A. Brief introduction per PET
8. Appendix B. Security scenarios
9. Appendix C. Attack vectors

Federated Analytics/Learning

In the tree, we use the term Federated Analytics to describe statistical analysis (e.g., aggregation) that is performed in some federated manner. We use Federated Learning to signify specifically that the process of model training occurs in a federated manner.