Differential privacy reduces the effective tail sample size in CVaR optimization to ε n τ, yielding explicit excess risk rates that separate statistical tail error from a privacy price term.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
The Privacy Price of Tail-Risk Learning: Effective Tail Sample Size in Differentially Private CVaR Optimization
Differential privacy reduces the effective tail sample size in CVaR optimization to ε n τ, yielding explicit excess risk rates that separate statistical tail error from a privacy price term.