Distributed mini-batch random projection algorithms for reduced communication overhead

Author

Soomin Lee ; Nedic, Angelia

Author_Institution

Electr. & Comput. Eng, Univ. of Illinois, Urbana, IL, USA

fYear

2013

fDate

3-5 Dec. 2013

Firstpage

559

Lastpage

562

Abstract

We propose a gossip-based mini-batch random projection (GMRP) algorithm that can reduce communication overhead for a distributed optimization problem defined over a network with a very large number of constraints. We state a convergence result and provide an application of the GMRP, text classification with support vector machines.

Keywords

convergence; distributed algorithms; optimisation; randomised algorithms; GMRP algorithm; convergence result; distributed mini-batch random projection algorithms; distributed optimization problem; gossip-based mini-batch random projection algorithm; reduced communication overhead; support vector machines; text classification; Convergence; Educational institutions; Optimization; Projection algorithms; Support vector machines; Training; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Global Conference on Signal and Information Processing (GlobalSIP), 2013 IEEE

Conference_Location

Austin, TX

Type

conf

DOI

10.1109/GlobalSIP.2013.6736939

Filename

6736939

Link To Document

https://search.isc.ac/dl/search/defaultta.aspx?DTC=49&DC=3254183