Reproduce “Task-aware Kernel Stein certificates for Rényi differential privacy” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Leoni Carla Wirth, Tara Trauthwein, Gesine Reinert
The authors develop task-aware Rényi Pufferfish privacy certificates using kernel Stein discrepancies (KSDs) matched to secret comparisons and output directions. An escort KSD characterizes the maximum first-order change in Rényi divergence over admitted perturbations; global bounds require conditions. In an ACSIncome illustration, a projected Gaussian release has a lower privacy budget than the full release, only for the projected output.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.