Finite-Time Analysis of Stratified Sampling for Monte Carlo

Alexandra Carpentier 1 Rémi Munos 1
1 SEQUEL - Sequential Learning
LIFL - Laboratoire d'Informatique Fondamentale de Lille, Inria Lille - Nord Europe, LAGIS - Laboratoire d'Automatique, Génie Informatique et Signal
Abstract : We consider the problem of stratified sampling for Monte-Carlo integration. We model this problem in a multi-armed bandit setting, where the arms represent the strata, and the goal is to estimate a weighted average of the mean values of the arms. We propose a strategy that samples the arms according to an upper bound on their standard deviations and compare its estimation quality to an ideal allocation that would know the standard deviations of the strata. We provide two regret analyses: a distribution-dependent bound $\widetilde O(n^{-3/2})$ that depends on a measure of the disparity of the strata, and a distribution-free bound $\widetilde O(n^{-4/3})$ that does not.
Document type :
Conference papers
Complete list of metadatas

Cited literature [15 references]  Display  Hide  Download

https://hal.inria.fr/inria-00636924
Contributor : Alexandra Carpentier <>
Submitted on : Monday, February 27, 2012 - 5:46:09 PM
Last modification on : Thursday, February 21, 2019 - 10:52:49 AM
Long-term archiving on: Thursday, June 14, 2012 - 5:00:24 PM

File

mc-ucb_3.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : inria-00636924, version 3

Collections

Citation

Alexandra Carpentier, Rémi Munos. Finite-Time Analysis of Stratified Sampling for Monte Carlo. NIPS - Twenty-Fifth Annual Conference on Neural Information Processing Systems, Dec 2011, Grenade, Spain. ⟨inria-00636924v3⟩

Share

Metrics

Record views

509

Files downloads

178