A parallel decomposition solver for SVM: Distributed dual ascend using fenchel duality

Tamir Hazan*, Amit Man, Amnon Shashua

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

23 Scopus citations

Abstract

We introduce a distributed algorithm for solving large scale Support Vector Machines (SVM) problems. The algorithm divides the training set into a number of processing nodes each running independently an SVM sub-problem associated with its subset of training data. The algorithm is a parallel (Jacobi) block-update scheme derived from the convex conjugate (Fenchel Duality) form of the original SVM problem. Each update step consists of a modified SVM solver running in parallel over the sub-problems followed by a simple global update. We derive bounds on the number of updates showing that the number of iterations (independent SVM applications on sub-problems) required to obtain a solution of accuracy ε is O(log(1/ε)). We demonstrate the efficiency and applicability of our algorithms by running on large scale experiments on standardized datasets while comparing the results to the state-of-the-art SVM solvers.

Original languageAmerican English
Title of host publication26th IEEE Conference on Computer Vision and Pattern Recognition, CVPR
DOIs
StatePublished - 2008
Event26th IEEE Conference on Computer Vision and Pattern Recognition, CVPR - Anchorage, AK, United States
Duration: 23 Jun 200828 Jun 2008

Publication series

Name26th IEEE Conference on Computer Vision and Pattern Recognition, CVPR

Conference

Conference26th IEEE Conference on Computer Vision and Pattern Recognition, CVPR
Country/TerritoryUnited States
CityAnchorage, AK
Period23/06/0828/06/08

Fingerprint

Dive into the research topics of 'A parallel decomposition solver for SVM: Distributed dual ascend using fenchel duality'. Together they form a unique fingerprint.

Cite this