Overview
Explore the complexities of two-party differential privacy in this 40-minute lecture from the Simons Institute. Delve into the challenges of distributed differential privacy, where parties analyze joint data while preserving privacy for both datasets. Examine the accuracy gap between distributed solutions and client-server settings for fundamental functions like inner product and Hamming distance. Understand the inherent limitations proven by McGregor et al. and learn how computational differential privacy can bypass these constraints using public-key cryptography. Discover new research proving the necessity of public-key cryptography in overcoming these limitations, with implications for key-agreement protocols. Investigate the connection between non-Boolean inner product of independent Santha-Vazirani sources and good condensers. Gain insights into the inner product of a single, strong SV source with a uniformly random seed as a good condenser, even with dependency between seed and source. Join Naom Mazor from UC Berkeley as he presents joint work with Iftach Haitner, Jad Silbak, and Eliad Tsfadia in this illuminating talk on minimal complexity assumptions for cryptography.
Syllabus
On the Complexity of Two-Party Differential Privacy
Taught by
Simons Institute