AI Paper Atlas

Research map / Federated learning and privacy / Differential privacy

Learning-augmented private algorithms for multiple quantile release

Mikhail Khodak, Kareem Amin, Travis Dick, Sergei Vassilvitskii

ICML 2023 · Federated learning and privacy › Differential privacy · approach: Privacy accounting and mechanisms · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use K-ary AQ with edge-based prior adaptation instead of regular binary AQMultiple quantile release · Gaussian and Adult age datasetsmaximum quantile gapMixed
Fit Laplace priors to public data with PubFit rather than use fixed priorsPublic-private quantile release · Adult age and hoursquantile release errorImproved
Use PubFit rather than public quantiles under distribution shiftPublic-private quantile release · Goodreads ratingsquantile release errorImproved
Mix PubFit's learned priors with a half-Cauchy prior for robustness (ablation)PubFit · Goodreads page counts under distribution shiftquantile release errorImproved
Use PubProx rather than the Uniform (AQ) baselineSequential quantile release · Synthetic and CitiBiketime-averaged maximum quantile gapImproved
Use PubProx rather than PubPrev with the same informationSequential quantile release · Synthetic, CitiBike, and BBCtime-averaged quantile release errorImproved
Mix learned PubProx priors with a Cauchy or half-Cauchy prior using coefficient 0.1 (ablation)PubProx · Sequential release tasksquantile release errorImproved
Mix PubPrev priors with a Cauchy or half-Cauchy prior using coefficient 0.1 (ablation)PubPrev · CitiBikequantile release errorImproved
Use a static Cauchy prior rather than PubProxSequential quantile release · BBC, mean performancemean quantile release errorImproved
Use PubProx rather than a static Cauchy priorSequential quantile release · BBC, typical headlinemedian quantile release errorImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy