• Cases where privacy concerns are not “local” (e.g. privacy for large groups) or utility is not “global” (e.g. targeting) • Matching guarantees with privacy law & regulation
Di erential privacy gives a way to analyze data that provably doesn't leak (much) information about individuals. How can we be sure that the predictions/decisions treat di erent groups fairly? What does this even mean?
thinking about privacy being noticeable in a huge range of disciplines, ranging from traditional areas of computer science (databases, machine learning, networking, security) to economics and game theory, false discovery control, o cial statistics and
DP protects individual against ALL harms regardless of prior knowledge. Fun paper: “Is Terry Gross protected?” Harm from the result itself cannot be eliminated. DP makes it impossible to guess whether one participated in a database with large probability. Only true under assumption that there is no group structure.
Differential privacy in the real world | PPT - SlideShare
2018年11月14日 · This document discusses privacy-preserving artificial intelligence and differential privacy. Differential privacy is a technique that adds mathematical noise to aggregate data to prove that an individual's private data is protected.
By calibrating the noise with the sensitivity of the function and the privacy parameter epsilon, we ensure that the output maintains epsilon differential privacy. So for evaluating a count, we calculate the exact count and then add to it noise distributed Laplace one over epsilon, and for evaluating a sum, we calculate the exact sum, and then ...
What is Differential Privacy? | PPT - SlideShare
2018年3月28日 · Differential privacy is mathematical definition for the privacy loss that results to individuals when their private information is used to create an AI product. It can be used to build customer trust, making those customers more likely to share their data with you.
- [PDF]
Differential Privacy
2021年10月5日 · “Statistical” Privacy for All Computations Differential privacy preserves “I could have opted out” privacy for every computation, including total population counts Image credit: Keydifferences.com
A randomized function K gives ε-differential privacy if for all data sets D1 and D2 differing on at most one element, and all S Range(K), Pr[K(D1) in S] ≤ exp(ε) x Pr[K(D2) in S]
Johnny’s mom may now be able tell if he watches R-rated movies from A(G). What does G~G’ mean? Example: Add/delete one row? -differential privacy. Why does this help? ...