Non-linearly increasing resampling in racing algorithms

Verena Heidrich-Meisner 1 Christian Igel 2
1 TAO - Machine Learning and Optimisation
LRI - Laboratoire de Recherche en Informatique, UP11 - Université Paris-Sud - Paris 11, Inria Saclay - Ile de France, CNRS - Centre National de la Recherche Scientifique : UMR8623
Abstract : Racing algorithms are iterative methods for identifying the best among several options with high probability. The quality of each option is a random variable. It is estimated by its empirical mean and concentration bounds obtained from repeated sampling. In each iteration of a standard racing algorithm each promising option is reevaluated once before being statistically compared with its competitors. We argue that Hoeffding and empirical Bernstein races benefit from generalizing the functional dependence of the racing iteration and the number of samples per option and illustrate this on an artificial benchmark problem.
Type de document :
Communication dans un congrès
Michel Verleysen. European Symposium on Artificial Neural Networks, Apr 2011, Bruges, Belgium. Evere, Belgium: d-side publications, pp.465-470, 2011, 19th European Symposium on Artificial Neural Networks (ESANN 2011). 〈http://www.dice.ucl.ac.be/Proceedings/esann/esannpdf/es2011-75.pdf〉
Liste complète des métadonnées

Littérature citée [8 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/inria-00633006
Contributeur : Verena Heidrich-Meisner <>
Soumis le : lundi 17 octobre 2011 - 11:59:10
Dernière modification le : jeudi 5 avril 2018 - 12:30:12
Document(s) archivé(s) le : jeudi 15 novembre 2012 - 09:50:09

Fichier

esann2011.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : inria-00633006, version 1

Collections

Citation

Verena Heidrich-Meisner, Christian Igel. Non-linearly increasing resampling in racing algorithms. Michel Verleysen. European Symposium on Artificial Neural Networks, Apr 2011, Bruges, Belgium. Evere, Belgium: d-side publications, pp.465-470, 2011, 19th European Symposium on Artificial Neural Networks (ESANN 2011). 〈http://www.dice.ucl.ac.be/Proceedings/esann/esannpdf/es2011-75.pdf〉. 〈inria-00633006〉

Partager

Métriques

Consultations de la notice

395

Téléchargements de fichiers

142