Accelerated directional search with non-Euclidean prox-structure

DOI10.1134/S0005117919040076zbMath1434.90143arXiv1710.00162OpenAlexW2796630182MaRDI QIDQ2290400

É. A. Gorbunov, Evgeniya A. Vorontsova, Alexander V. Gasnikov

Publication date: 27 January 2020

Published in: Automation and Remote Control (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1710.00162

zbMATH Keywords

convex optimizationnon-Euclidean prox-structurefirst-order accelerated methodslinear coupling methoduniform measure concentration on single euclidean sphere

Mathematics Subject Classification ID

Convex programming (90C25) Derivative-free methods and methods using generalized derivatives (90C56)

Related Items

First-order methods for convex optimization ⋮ On the upper bound for the expectation of the norm of a vector uniformly distributed on the sphere and the phenomenon of concentration of uniform measure on the sphere ⋮ Accelerated gradient-free optimization methods with a non-Euclidean proximal operator ⋮ ACDS

Uses Software

• DiffSharp
• GitHub

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• On lower complexity bounds for large-scale smooth convex optimization
• Gradient-free proximal methods with inexact oracle for convex stochastic nonsmooth optimization problems on the simplex
• Gradient-free two-point methods for solving stochastic nonsmooth convex optimization problems with small non-random noises
• Universal method for stochastic composite optimization problems
• On the upper bound for the expectation of the norm of a vector uniformly distributed on the sphere and the phenomenon of concentration of uniform measure on the sphere
• Stochastic online optimization. Single-point and multi-point non-linear multi-armed bandits. Convex and strongly-convex case
• Random gradient-free minimization of convex functions
• Efficiency of Coordinate Descent Methods on Huge-Scale Optimization Problems
• Linear Coupling: An Ultimate Unification of Gradient and Mirror Descent