Differentially Private Sampling from Distributions

Explore the intricacies of differentially private sampling from distributions in this 30-minute lecture by Satchit Sivakumar from Boston University. Delve into the importance of private sampling, its accuracy, and the context surrounding this field of study. Gain insights into the speaker's contributions and related work in the area. Examine various techniques, including a simple example of Bernoulli with bounded bias, and understand the privacy implications. Follow along as the speaker provides an overview of product lower bound and presents a visual depiction of the algorithm. Conclude with a comprehensive summary of the key takeaways from this talk, which is part of the "Workshop on Differential Privacy and Statistical Data Analysis" hosted by the Fields Institute.

Intro
Private Sampling
Sampling Accuracy
Context
Why study this problem?
Summary of Contributions
Related Work
Techniques
Simple Example: Bernoulli with Bounded 57
Bernoulli with Bounded Bias: Privacy
Overview of product lower bound
Overview of proof (fairytale version)
Visual depiction of algorithm
Conclusion