With social media and big data everywhere, data privacy has been a growing, public concern. Recognizing this issue, entities such as Google, Apple, and the US Census Bureau are promoting better privacy techniques; specifically differential privacy, a mathematical condition that quantifies privacy risk. In this course, you will learn to code basic data privacy methods and a differentially private algorithm based on various differentially private properties. With these tools in hand, you will learn how to generate a basic synthetic (fake) data set with the differential privacy guarantee for public data release.
Introduction to Data PrivacyFree
This chapter covers some basic data privacy techniques that statisticians use to anonymize data. You'll first learn how to remove identifiers and then generate synthetic data from probability distributions.
Introduction to Differential Privacy
After covering the basic data privacy techniques, you'll learn conceptually about differential privacy as well as how to implement the most popular and common differentially private algorithm called the Laplace mechanism.
Differentially Private Properties
In this chapter, you will learn the various properties of differential privacy, such as the combination rules and post-processing, to properly implement the Laplace mechanism for various kinds data questions.
Differentially Private Data Synthesis
In this chapter, you will learn how to release simple data sets publicly using differentially private data synthesis techniques.
Postdoctoral Researcher at the Los Alamos National Laboratory
Claire McKay Bowen is a Postdoctoral Researcher in the Statistical Science Group at the Los Alamos National Laboratory. She conducts research in uncertainty quantification with physics-informed Bayesian Model updating and data privacy via differentially private data synthesis methods. Her other interests include statistical computing, scientific communication, and STEM outreach.