Utility Analysis for Differentially-Private Pipelines

Explore a 13-minute conference talk from PEPR '24 that delves into utility analysis for differentially-private pipelines. Learn about a methodology and open-source module within the PipelineDP framework designed to address the challenges of implementing differential privacy. Discover how this approach estimates the impact of differential privacy techniques on data quality and aids in selecting optimal hyperparameters. Gain insights into the complex trade-offs between privacy guarantees and data quality, including the effects of aggregation, outlier handling, and noise addition. Understand how this user-friendly tool can assist in making informed decisions when implementing differentially private pipelines in Python, ultimately enhancing the usability of differential privacy tools.