2021 (modified: 15 May 2022)ICML 2021Readers: Everyone
Abstract:What guarantees are possible for solving logistic regression in one pass over a data stream? To answer this question, we present the first data oblivious sketch for logistic regression. Our sketch ...