A GDP-based auditing framework delivers the first tight privacy audits for MST and AIM, with empirical μ of 0.43 close to the implied 0.45 for (ε,δ)=(1,10^{-2}).
2018 Differential Privacy Synthetic Data Challenge
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Tight Auditing of Differential Privacy in MST and AIM
A GDP-based auditing framework delivers the first tight privacy audits for MST and AIM, with empirical μ of 0.43 close to the implied 0.45 for (ε,δ)=(1,10^{-2}).