Solving Ridge Regression using Sketched Preconditioned SVRG
2016-02-07Unverified0· sign in to hype
Alon Gonen, Francesco Orabona, Shai Shalev-Shwartz
Unverified — Be the first to reproduce this paper.
ReproduceAbstract
We develop a novel preconditioning method for ridge regression, based on recent linear sketching methods. By equipping Stochastic Variance Reduced Gradient (SVRG) with this preconditioning process, we obtain a significant speed-up relative to fast stochastic methods such as SVRG, SDCA and SAG.