A linear-time algorithm using robust statistics achieves near-optimal excess risk for user-level private convex optimization under ℓ1/ℓ∞ geometry, up to an extra factor of ε.
User-level differentially private stochastic convex optimization: Efficient algorithms with optimal rates
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Linear-Time User-Level DP-SCO via Robust Statistics
A linear-time algorithm using robust statistics achieves near-optimal excess risk for user-level private convex optimization under ℓ1/ℓ∞ geometry, up to an extra factor of ε.