Non-linearly increasing resampling in racing algorithms

V. Heidrich-Meisner; Christian Igel

Non-linearly increasing resampling in racing algorithms

V. Heidrich-Meisner, Christian Igel

Department of Computer Science

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.

Original language	English
Title of host publication	19th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2011)
Editors	M. Verleysen
Number of pages	6
Publication date	2011
Pages	465-470
ISBN (Print)	978-2-87419-044-5
Publication status	Published - 2011
Event	19th European Symposium On Artificial Neural Networks, Computational Intelligence and Machine Learning - Bruges, Belgium Duration: 27 Apr 2011 → 27 Apr 2011 Conference number: 19

Conference

Conference	19th European Symposium On Artificial Neural Networks, Computational Intelligence and Machine Learning
Number	19
Country/Territory	Belgium
City	Bruges
Period	27/04/2011 → 27/04/2011

Access to Document

https://hal.inria.fr/inria-00633006/document

Cite this

Non-linearly increasing resampling in racing algorithms. / Heidrich-Meisner, V.; Igel, Christian.

19th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2011). ed. / M. Verleysen. 2011. p. 465-470.

Research output: Chapter in Book/Report/Conference proceeding › Conference abstract in proceedings › Research › peer-review

Heidrich-Meisner, V & Igel, C 2011, Non-linearly increasing resampling in racing algorithms. in M Verleysen (ed.), 19th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2011). pp. 465-470, 19th European Symposium On Artificial Neural Networks, Computational Intelligence and Machine Learning, Bruges, Belgium, 27/04/2011. <https://hal.inria.fr/inria-00633006/document>

@inbook{4d516b668dfd410386e2f673907d5ae9,

title = "Non-linearly increasing resampling in racing algorithms",

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.",

author = "V. Heidrich-Meisner and Christian Igel",

year = "2011",

language = "English",

isbn = "978-2-87419-044-5",

pages = "465--470",

editor = "M. Verleysen",

booktitle = "19th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2011)",

note = "19th European Symposium On Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2011 ; Conference date: 27-04-2011 Through 27-04-2011",

}

TY - ABST

T1 - Non-linearly increasing resampling in racing algorithms

AU - Heidrich-Meisner, V.

AU - Igel, Christian

N1 - Conference code: 19

PY - 2011

Y1 - 2011

N2 - 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.

AB - 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.

M3 - Conference abstract in proceedings

SN - 978-2-87419-044-5

SP - 465

EP - 470

BT - 19th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2011)

A2 - Verleysen, M.

T2 - 19th European Symposium On Artificial Neural Networks, Computational Intelligence and Machine Learning

Y2 - 27 April 2011 through 27 April 2011

ER -

Non-linearly increasing resampling in racing algorithms

Abstract

Conference

Access to Document

Fingerprint

Cite this