Energy Differential Privacy (EDP)

Energy Differential Privacy (EDP) enables the use of the gold standard of privacy protection, differential privacy, for high value energy efficiency analytics.

This project is funded by the U.S. Department of Energy as part of the Energy Data Vault project.

This documentation covers the core library (eeprivacy) as well as detailed examples of applying the library in practice.

In the ecosystem of open source differential privacy projects, EDP targets a sector-specific niche, enabling and optimizing specific use cases rather than solving general purpose problems.

Core Library

• eeprivacy
  • Operations
  • Mechanisms
    • A Note on Confidence Intervals

Examples

• Laplace Mechanism Basics
  • Distribution of Laplace Mechanism Outputs
  • Laplace Mechanism Confidence Interval
  • Examples of Laplace Mechanism Helpers using eeprivacy
  • Computing Private Histograms
• Gaussian Mechanism Basics
  • Distribution of Gaussian Mechanism Outputs
  • Gaussian Mechanism Confidence Interval
• Stochastically Testing Privacy Mechanisms
• Computing Bounds for Clamped Means
  • Dataset
  • Background Theory
  • Approach to Privately Finding Bounds
  • Private Quantiles
  • Private Quantile Accuracy
  • Results
• Computing Clamped Mean Bounds with the Sparse Vector Technique
  • Sparse Vector Technique for Clamping Bounds
• Private Load Shape Algorithm Design
  • A Semantic Model for Load Shapes
  • Candidate Mechanisms
  • Sample Data Set
  • Accuracy/Privacy Trade-Off of Approaches
• Computing Private Load Shapes
  • Dataset
  • Applying Gaussian Mechanism for Load Shapes

Applications

• Private Comparison Groups
  • Dataset
  • Population Mean with Laplace Mechanism
  • Population P99 with Laplace Mechanism
  • Comparison Group Percent Change in Consumption with Laplace Mechanism
  • Comparison Group Average Load Shape with Gaussian Mechanism

Indices and tables

• Index

• Module Index

• Search Page