LGFeb 7, 2016
Solving Ridge Regression using Sketched Preconditioned SVRG
arXiv:1602.02350v249 citations
Originality Incremental advance
AI Analysis
This work addresses computational efficiency in ridge regression for machine learning practitioners, representing an incremental improvement over existing stochastic optimization methods.
The paper tackled ridge regression by developing a novel preconditioning method using linear sketching, which when combined with SVRG achieved a significant speed-up compared to existing fast stochastic methods like SVRG, SDCA, and SAG.
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.