| HAL : hal-00696072, version 1 |
| arXiv : 1205.2282 |
| Fiche détaillée | Récupérer au format |
|
|
| 20-th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012), Bruges : Belgique (2012) |
|
|
|
|
| A Discussion on Parallelization Schemes for Stochastic Vector Quantization Algorithms |
|
|
| Matthieu Durut 1, 2Benoît Patra 2, 3 |
|
|
| (04/2012) |
|
|
| 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. |
|
|
|
|
|
|
|
|
|
|
| 1 : | Laboratoire Traitement et Communication de l'Information [Paris] (LTCI) |
| Télécom ParisTech – CNRS : UMR5141 | |
| 2 : | Lokad |
| Lokad | |
| 3 : | Laboratoire de Statistique Théorique et Appliquée (LSTA) |
| Université Pierre et Marie Curie (UPMC) - Paris VI | |
| 4 : | Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne) (SAMM) |
| Université Paris I - Panthéon-Sorbonne | |
|
|
|
|
|
|
|
|
| Domaine | : | Statistiques/Machine Learning Informatique/Apprentissage Informatique/Calcul parallèle, distribué et partagé |
|
|
| Liste des fichiers attachés à ce document : | |||||||||||||
|
|
|
| hal-00696072, version 1 | |
| http://hal.archives-ouvertes.fr/hal-00696072 | |
| oai:hal.archives-ouvertes.fr:hal-00696072 | |
| Contributeur : Fabrice Rossi | |
| Soumis le : Jeudi 10 Mai 2012, 16:26:12 | |
| Dernière modification le : Jeudi 10 Mai 2012, 16:44:35 | |