First-order Methods for Nonsmooth Convex Large-scale Optimization, I: General Purpose Methods

By Anatoli B. Juditsky, Anatoli Juditsky@imag Fr

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

A Universal Catalyst for First-Order Optimization (2015, methodology): “Such methods should be contrasted with stochastic gradient methods [15, 12], which minimize a different non-deterministic function.”
Incremental Majorization-Minimization Optimization with Application to Large-Scale Machine Learning (2014, background): “Another popular algorithm is the stochastic mirror descent algorithm (see [32]) for general nonsmooth convex problems, a setting we do not consider in this paper since nonsmooth functions do not always admit practical first-order surrogates.”
On stochastic proximal gradient algorithms (2014, unknown): “23 65 v1 [ m at h. ST ] 1 1 Fe b 20”
Optimal Primal-Dual Methods for a Class of Saddle Point Problems (2013, methodology): “, by applying an improved algorithm in [20].”
On Perturbed Proximal Gradient Algorithms (2014, background): “Some of these works focused on stochastic sub-gradient and mirror descent algorithms; see Nemirovski et al. (2008); Duchi et al. (2011); Cotter et al. (2011); Lan (2012); Juditsky and Nemirovski (2012a,b).”
Optimal Stochastic Approximation Algorithms for Strongly Convex Stochastic Composite Optimization I: A Generic Algorithmic Framework (2012, methodology): “This is the standard setting used in the regular SA methods [23, 14].”
Proximal Reinforcement Learning: A New Theory of Sequential Decision Making in Primal-Dual Spaces (2014, methodology): “Nash games are closely related to saddle point problems [129, 130, 135].”