We develop a new tool for data-dependent analysis of the exploration-exploitation trade-off in learning under limited feedback. Our tool is based on two main ingredients. The first ingredient is a new concentration inequality that makes it possible to control the concentration of weighted averages of multiple (possibly uncountably many) simultaneously evolving and interdependent martingales. The second ingredient is an application of this inequality to the exploration-exploitation trade-off via importance weighted sampling. We apply the new tool to the stochastic multiarmed bandit problem, however, the main importance of this paper is the development and understanding of the new tool rather than improvement of existing algorithms for stochastic multiarmed bandits. In the follow-up work we demonstrate that the new tool can improve over state-of-the-art in structurally richer problems, such as stochastic multiarmed bandits with side information.

PAC-Bayes-Bernstein inequality for martingales and its application to multiarmed bandits / Y. Seldin, N. Cesa-Bianchi, P. Auer, F. Laviolette, J. Shawe Taylor - In: Proceedings of the Workshop on on-line trading of exploration and exploitation 2 : july 2, 2011, Bellevue, Washington, USA[s.l] : Microtome, 2012. (( convegno On-line Trading of Exploration and Exploitation tenutosi a Bellevue, Washington, USA nel 2011.

PAC-Bayes-Bernstein inequality for martingales and its application to multiarmed bandits

N. Cesa-Bianchi
Secondo
;
2012

Abstract

We develop a new tool for data-dependent analysis of the exploration-exploitation trade-off in learning under limited feedback. Our tool is based on two main ingredients. The first ingredient is a new concentration inequality that makes it possible to control the concentration of weighted averages of multiple (possibly uncountably many) simultaneously evolving and interdependent martingales. The second ingredient is an application of this inequality to the exploration-exploitation trade-off via importance weighted sampling. We apply the new tool to the stochastic multiarmed bandit problem, however, the main importance of this paper is the development and understanding of the new tool rather than improvement of existing algorithms for stochastic multiarmed bandits. In the follow-up work we demonstrate that the new tool can improve over state-of-the-art in structurally richer problems, such as stochastic multiarmed bandits with side information.
PAC-Bayesian analysis ; Bernstein's inequality ; Martingales ; Multiarmed bandits ; Model order selection ; Exploration-exploitation trade-off
Settore INF/01 - Informatica
   Pattern Analysis, Statistical Modelling and Computational Learning 2
   PASCAL2
   EUROPEAN COMMISSION
   FP7
   216886
2012
Book Part (author)
File in questo prodotto:
File Dimensione Formato  
seldin12a.pdf

accesso aperto

Tipologia: Post-print, accepted manuscript ecc. (versione accettata dall'editore)
Dimensione 428.19 kB
Formato Adobe PDF
428.19 kB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2434/179282
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact