may en lieunghien nguy - ezgrejestracja.pl
Huy Lê Nguy˜ên Eric Xu September 3, 2020 arXiv:2009 ... Theorem 1.1. There is an (ǫp,δp) differentially private algorithm for the k-means problem that achieves a utility bound of O(1)fD(OPTD) +O k∆ 2log nlog1/δp ǫp +O ∆2k p dlog1/δp ǫp where D is the input dataset, fD(OPTD) is the optimal k-means cost for the input dataset D, d is the ambient dimension of D, n is the cardinality ...
Đọc thêm