Reproduce “Robust and Differentially Private Mean Estimation” 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
Xiyang Liu, Weihao Kong, Sham Kakade, Sewoong Oh
PRIME combines privacy and adaptive filtering to estimate a mean despite differential privacy and arbitrary corruption. It uses matrix multiplicative weights to score points across directions and a private threshold to remove suspicious points. Under stated sub-Gaussian assumptions, it achieves error O(α√log(1/α)) with probability 0.9 in polynomial time, but requires an extra √d in its privacy-related sample term compared with an exponential-time method.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.