PostRI computes randomization intervals after differentially private median estimation, delivering 14-850% higher median utility than prior methods while keeping narrow intervals.
InThe Thirty Seventh Annual Conference on Learning Theory
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
-
Interpreting the Error of Differentially Private Median Queries through Randomization Intervals
PostRI computes randomization intervals after differentially private median estimation, delivering 14-850% higher median utility than prior methods while keeping narrow intervals.