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

Download:

Current browse context:

math

Change to browse by:

References & Citations

Bookmark

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

Mathematics > Optimization and Control

Title: One-Point Gradient-Free Methods for Composite Optimization with Applications to Distributed Optimization

Abstract: This work is devoted to solving the composite optimization problem with the mixture oracle: for the smooth part of the problem, we have access to the gradient, and for the non-smooth part, only to the one-point zero-order oracle. For such a setup, we present a new method based on the sliding algorithm. Our method allows to separate the oracle complexities and compute the gradient for one of the function as rarely as possible. The paper also present the applicability of our new method to the problems of distributed optimization and federated learning. Experimental results confirm the theory.
Comments: New in v2: completely new text of the paper; 26 pages, 1 figure, 2 tables, 1 algorithm
Subjects: Optimization and Control (math.OC); Distributed, Parallel, and Cluster Computing (cs.DC)
Cite as: arXiv:2107.05951 [math.OC]
  (or arXiv:2107.05951v2 [math.OC] for this version)

Submission history

From: Aleksandr Beznosikov [view email]
[v1] Tue, 13 Jul 2021 09:51:04 GMT (26kb)
[v2] Wed, 16 Feb 2022 20:40:56 GMT (618kb,D)

Link back to: arXiv, form interface, contact.