<ul><li>For scalable machine learning on large data sets, importance sampling is commonly used to subsample a representative subset for efficient model training.</li><li>This paper examines the privacy properties of importance sampling, specifically focusing on individualized privacy analysis.</li><li>The study finds that privacy in importance sampling is aligned with utility but conflicts with sample size.</li><li>The paper proposes two approaches for constructing sampling distributions that optimize privacy-efficiency trade-off and provide utility guarantees through coresets.</li></ul>

Personalized Privacy Amplification via Importance Sampling

Discover more