AI Paper Atlas

Research map / Federated learning and privacy / Privacy attacks and secure learning

A General Framework for Auditing Differentially Private Machine Learning

Fred Lu, Joseph Munoz, Maya Fuchs, Tyler LeBlond, Elliott V. Zaresky-Williams, Edward Raff, Francis Ferraro, Brian Testa

NeurIPS 2022 · Federated learning and privacy › Privacy attacks and secure learning · approach: Membership inference and auditing · 6 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
influence-based perturbation attack instead of ClipBKDML-Audit · logistic regression, Naive Bayes, random forestdetected privacy lower boundImproved
influence-based attack instead of Swap-X baselineML-Audit · iris, breast-cancer, output-perturbed logistic regressiondetected privacy lower boundNo clear effect
target class prior (flip label) instead of only class mean (ablation)Naive Bayes attack · Naive Bayesdetected privacy lower boundImproved
use Katz-log interval instead of Clopper-Pearson (ablation)ML-Audit · coverage simulationconfidence interval tightnessImproved
increase number of perturbed copies k at small eps (ablation)ML-Audit · auditing at small epsdetected privacy lower boundImproved
increase number of perturbed copies k at large eps (ablation)ML-Audit · auditing at large epsdetected privacy lower boundHurt
add exposed class counts to summary tauML-Audit · Naive Bayes (diffprivlib)detected privacy lower boundImproved
add zero-vector and test predictions to tau (ablation)ClipBKD · DP-SGD FMNISTdetected privacy lower boundImproved
normalize FMNIST to [-1,1] or [0,1] instead of [0,255] (ablation)DP-SGD audit · FMNISTdetected privacy risk: 0.7 -> 0.4Hurt
shadow adversarial attack instead of ClipBKD (ablation)DP-SGD audit · FMNIST vs CIFAR10detected privacy lower boundMixed

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