Abstract:
We propose a new way of justifying the accelerated gradient sliding of G. Lan, which allows one to extend the sliding technique to a combination of an accelerated gradient method with an accelerated variance reduction method. New optimal estimates for the solution of the problem of minimizing a sum of smooth strongly convex functions with a smooth regularizer are obtained.
Keywords:
accelerated gradient sliding of G. Lan, accelerated variance reduction methods, smooth strongly convex functions.
This work was supported by the Russian Foundation for Basic Research, project no. 18-31-20005 mol_a_ved (Section 1) and project no. 19-31-90062 Graduate students (Section 2).
Dvinskikh acknowledges the support of the Ministry of Science and Higher Education of the Russian Federation, state assignment no. 075-00337-20-03.
Presented:Yu. G. Evtushenko Received: 20.03.2020 Revised: 26.03.2020 Accepted: 03.04.2020
Citation:
D. M. Dvinskikh, S. S. Omelchenko, A. V. Gasnikov, A. I. Turin, “Accelerated gradient sliding for minimizing a sum of functions”, Dokl. RAN. Math. Inf. Proc. Upr., 492 (2020), 85–88; Dokl. Math., 101:3 (2020), 244–246