Reproduce “Complexity of Differentially Private Selection via Federated APIs” 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
Hilal Asi, Vitaly Feldman, Jelani Nelson, Huy L. Nguyen, Kunal Talwar, Samson Zhou
The authors introduce a differential-privacy threshold primitive based on the Sparse Vector Technique (SVT) to select among d candidates with bounded per-user scores. It checks whether a candidate’s summed score clears a threshold, without additional privacy cost for repeated negative answers. Under stated conditions, selection error is O(log(dn)/ε), versus an aggregation-only lower bound of order √d/ε under specified access and privacy assumptions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.