Laboratory of Structural Methods of Data Analysis in Predictive
Modeling Moscow Institute of Physics and Technology
Stochastic Intermediate Gradient Method for Convex Problems with Inexact Stochastic Oracle
In this paper we propose new method for convex optimization problems with inexact stochastic oracle. This method is an extension of the intermediate gradient method proposed by O. Devolder, F. Glineur, and Yu. Nesterov for problems with inexact oracle.
Our new method can be applied to the problems with composite structure, stochastic inexact oracle and allows using non-Euclidean setup. Also it allows to play on the tradeoff between the rate of convergence and oracle error accumulation depending on the problem parameters.

Авторы: Dvurechensky Pavel , Gasnikov Alexander

Дата: 27 декабря 2014

Статус: опубликована

Журнал: Сборник трудов 38-й конференции-школы ИППИ РАН «Информационные технологии и системы - 2014»

Страницы: 386-392

Год: 2014

Google scholar:

Направления исследований

Structural optimization

Дополнительные материалы