hal-00696072, version 1

A Discussion on Parallelization Schemes for Stochastic Vector Quantization Algorithms

Matthieu Durut () ^1, 2, Benoît Patra () ^2, 3, Fabrice Rossi (, ) ⁴

20-th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012) (2012) 477-482

• 1:  Laboratoire Traitement et Communication de l'Information [Paris] (LTCI)
• http://www.ltci.telecom-paristech.fr/
  Télécom ParisTech – CNRS : UMR5141 CNRS LTCI Télécom ParisTech 46 rue Barrault F-75634 Paris Cedex 13 France
• 2:  Lokad
• http://www.lokad.com/
  Lokad 10 rue Philippe de Champaigne 75013 Paris France
• 3:  Laboratoire de Statistique Théorique et Appliquée (LSTA)
• http://www.lsta.upmc.fr
  Université Pierre et Marie Curie (UPMC) - Paris VI Université Pierre et Marie Curie (Paris 6) Tour 15-25 2-ième étage Boite courrier 158 4, place Jussieu 75252 Paris Cedex 05 France
• 4:  Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne) (SAMM)
• http://samm.univ-paris1.fr/
  Université Paris I - Panthéon-Sorbonne Centre Pierre Mendès France 90 Rue de Tolbiac - 75634 Paris Cedex 13 France

Bibliographic reference

• Type of document: Peer-reviewed conferences/proceedings
• Subject:

  Statistics/Machine Learning

  Computer Science/Learning

  Computer Science/Distributed, Parallel, and Cluster Computing

• Title: A Discussion on Parallelization Schemes for Stochastic Vector Quantization Algorithms
• Abstract: This paper studies parallelization schemes for stochastic Vector Quantization algorithms in order to obtain time speed-ups using distributed resources. We show that the most intuitive parallelization scheme does not lead to better performances than the sequential algorithm. Another distributed scheme is therefore introduced which obtains the expected speed-ups. Then, it is improved to fit implementation on distributed architectures where communications are slow and inter-machines synchronization too costly. The schemes are tested with simulated distributed architectures and, for the last one, with Microsoft Windows Azure platform obtaining speed-ups up to $32$ Virtual Machines.
• Fulltext language: English
• Production date: 2011-11
• Book title: Proceedings of the 20-th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)
• Audience: international
• Publication date: 2012-04
• Page, identifiant, ...: 477-482
• Conference or book title: 20-th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012)
• Conference date: 2012-04-25
• Conference date (end): 2012-04-27
• City: Bruges
• Country: Belgium

Attached file list to this document: 

TEX

ESANN_DAVQ_Last.aux(566 B)

ESANN_DAVQ_Last.bbl(876 B)

ESANN_DAVQ_Last.blg(204 B)

ESANN_DAVQ_Last.pdf(429.8 KB)

ESANN_DAVQ_Last.tex(16.8 KB)

correctedPa10NoTitle.pdf(44.9 KB)

dalvq_esann_2012.bib(1.3 KB)

esannV2.cls(21.3 KB)

pa10NoTitle.pdf(45 KB)

pa10RandomdelayedNoTitle.pdf(43.5 KB)

scalabilityNoTitle.pdf(51.9 KB)

ANNEX

spotlight-slide.pdf(27.6 KB)

poster.pdf(351.3 KB)

PDF

ESANN_DAVQ_Last.pdf(408.7 KB)

• hal-00696072, version 1
• http://hal.archives-ouvertes.fr/hal-00696072
• oai:hal.archives-ouvertes.fr:hal-00696072
• From: Fabrice Rossi <>
• Submitted on: Thursday, 10 May 2012 16:26:12
• Updated on: Thursday, 10 May 2012 16:44:35