Algorithms with gradient clipping for stochastic optimization with heavy-tailed noise

DOI10.1134/s1064562423701144MaRDI QIDQ6204268

Marina Danilova

Publication date: 27 March 2024

Published in: Doklady Mathematics (Search for Journal in Brave)

zbMATH Keywords

stochastic optimizationconvex optimizationfirst-order methods

Mathematics Subject Classification ID

Mathematical programming (90Cxx) Stochastic processes (60Gxx) Limit theorems in probability theory (60Fxx)

Cites Work

• First-order methods of smooth convex optimization with inexact oracle
• Finite-dimensional variational inequality and nonlinear complementarity problems: A survey of theory, algorithms and applications
• On tail probabilities for martingales
• Universal method for stochastic composite optimization problems
• On Bernstein-type inequalities for martingales.
• Algorithms of robust stochastic optimization based on mirror descent method
• Probability Inequalities for the Sum of Independent Random Variables
• Stochastic Model-Based Minimization of Weakly Convex Functions
• Optimal Stochastic Approximation Algorithms for Strongly Convex Stochastic Composite Optimization I: A Generic Algorithmic Framework
• Large-Scale Convex Optimization
• Stochastic First- and Zeroth-Order Methods for Nonconvex Stochastic Programming