Skip to yearly menu bar Skip to main content


Poster

Nonlinear Acceleration of Stochastic Algorithms

Damien Scieur · Francis Bach · Alexandre d'Aspremont

Pacific Ballroom #174

Keywords: [ Convex Optimization ] [ Stochastic Methods ]


Abstract:

Extrapolation methods use the last few iterates of an optimization algorithm to produce a better estimate of the optimum. They were shown to achieve optimal convergence rates in a deterministic setting using simple gradient iterates. Here, we study extrapolation methods in a stochastic setting, where the iterates are produced by either a simple or an accelerated stochastic gradient algorithm. We first derive convergence bounds for arbitrary, potentially biased perturbations, then produce asymptotic bounds using the ratio between the variance of the noise and the accuracy of the current point. Finally, we apply this acceleration technique to stochastic algorithms such as SGD, SAGA, SVRG and Katyusha in different settings, and show significant performance gains.

Live content is unavailable. Log in and register to view live content