Presentation 2008-12-19
On effect of balancing investment in nonstochastic multi-armed bandit problems
Taishi UCHIYA, Atsuyoshi NAKAMURA, Mineichi KUDO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The multiarmed bandit problem is a problem in which a gambler chooses one arm of K nonidentical slot machines to play in a sequence of trials so as to maximize his reward. Past solutions for the bandit problem have almost always relied on assumptions about the statistics of the slot machines. On the other hand, Auer et al. made no statistical assumption whatsoever about the nature of the process generating the payoffs of the slot machine. They gave solutions to the bandit ploblem in which an adversary has complate control over the payoffs. In this paper, we extend this problem to the problem of choosing more than one slot machine at a time and theoretically analyze it.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) multi-armed bandit problem / online learning
Paper # PRMU2008-183
Date of Issue

Conference Information
Committee PRMU
Conference Date 2008/12/11(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) On effect of balancing investment in nonstochastic multi-armed bandit problems
Sub Title (in English)
Keyword(1) multi-armed bandit problem
Keyword(2) online learning
1st Author's Name Taishi UCHIYA
1st Author's Affiliation Graduate School of Information Science and Technology Hokkaido University()
2nd Author's Name Atsuyoshi NAKAMURA
2nd Author's Affiliation Graduate School of Information Science and Technology Hokkaido University
3rd Author's Name Mineichi KUDO
3rd Author's Affiliation Graduate School of Information Science and Technology Hokkaido University
Date 2008-12-19
Paper # PRMU2008-183
Volume (vol) vol.108
Number (no) 363
Page pp.pp.-
#Pages 6
Date of Issue