<ul><li>Stochastic first-order methods for empirical risk minimization employ gradient approximations based on sampled data in lieu of exact gradients.</li><li>The recently proposed variance-reduction technique, alpha-SVRG, allows for fine-grained control of the level of residual noise in the learning dynamics, and has been reported to outperform SGD and SVRG in modern deep learning scenarios.</li><li>In strongly convex environments, alpha-SVRG has a faster convergence rate compared to SGD and SVRG under suitable choice of alpha.</li><li>Simulation results on linear regression validate the theory.</li></ul>

Convergence Analysis of alpha-SVRG under Strong Convexity

Discover more