We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat
< prev | next >
new | recent | 2201

Change to browse by:

cs
cs.LG
math
math.OC
stat.ML

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Mathematics > Optimization and Control

Title: On the Complexity of a Practical Primal-Dual Coordinate Method

Authors: Ahmet Alacaoglu, Volkan Cevher, Stephen J. Wright
(Submitted on 19 Jan 2022)
Abstract: We prove complexity bounds for the primal-dual algorithm with random extrapolation and coordinate descent (PURE-CD), which has been shown to obtain good practical performance for solving convex-concave min-max problems with bilinear coupling. Our complexity bounds either match or improve the best-known results in the literature for both dense and sparse (strongly)-convex-(strongly)-concave problems.
Subjects: Optimization and Control (math.OC); Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:2201.07684 [math.OC]
  (or arXiv:2201.07684v1 [math.OC] for this version)

Submission history

From: Ahmet Alacaoglu [view email]
[v1] Wed, 19 Jan 2022 16:14:27 GMT (43kb,D)
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.