AI Paper Atlas

Research map / Federated learning and privacy / Differential privacy

Renyi Differential Privacy Mechanisms for Posterior Sampling

Joseph Geumlek, Shuang Song, Kamalika Chaudhuri

NeurIPS 2017 · Federated learning and privacy › Differential privacy · approach: Private data analysis frameworks

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 the concentrated-prior posterior mechanism instead of OPSBayesian logistic regression · Abalone, Adult, MNIST 3vs8test errorImproved
Use the diffuse-likelihood posterior mechanism instead of the concentrated-prior mechanismBayesian logistic regression · Abalone, Adult, MNIST 3vs8test errorImproved
Use direct posterior sampling when its RDP guarantee permits it, rather than the Gaussian mechanismBeta-Bernoulli sampling · Synthetic dataKL divergence from the true posteriorImproved
Use concentrated-posterior sampling rather than diffuse-posterior samplingBeta-Bernoulli sampling · Synthetic data, lambda = 2KL divergence from the true posteriorMixed
Use prior-based posterior mechanisms rather than the Gaussian mechanismBeta-Bernoulli sampling · Synthetic data, lambda = 15KL divergence from the true posteriorMixed
Decrease the data-contribution scale rDiffuse Posterior · Beta-Bernoulli samplingRDP epsilonImproved
Decrease the prior scale m, strengthening the priorConcentrated Posterior · Beta-Bernoulli samplingRDP epsilonImproved
Use a more informative prior that better matches the populationBeta-Bernoulli prior-based mechanisms · Synthetic datalog likelihoodImproved

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