Article Source

• Differential Privacy And The Complexity Of Simple Queries

Differential Privacy And The Complexity Of Simple Queries

• https://simons.berkeley.edu/talks/differential-privacy-complexity-simple-queries
• Speaker: Jonathan Ullman (Northeastern University)

Abstract

Differential privacy has been a remarkably successful framework for studying individual privacy in statistical estimation and machine learning. While the information theoretic limits of differentially private statistical estimation are relatively well understood, our understanding of the computational complexity of differentially private estimation lags behind. In this talk I’ll survey some of the foundational results in differential privacy, discuss the known information-computation tradeoffs for differentially private statistical estimation, and highlight some important open problems.