PET Patterns

Patterns are reusable designs, not recipes. Each one states when to use it, when not to use it, what can still go wrong, and which research problems remain open.

Pattern Selection

Problem	Pattern
Aggregate measurement without centralizing raw data	Federated analytics
Cross-organization model training	Cross-silo federated learning
Safer data-like release	DP synthetic data release
Dataset overlap	Private set intersection
Model inference over protected inputs	Private inference
General-purpose protected inference	Confidential inference
Retrieval across sensitive corpora	Federated RAG
Fine-tuning on sensitive data	Private LLM fine-tuning