UCB revisited: improved regret bounds for the stochastic multi-armed bandit problem (Q653803)
From MaRDI portal
scientific article
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | UCB revisited: improved regret bounds for the stochastic multi-armed bandit problem |
scientific article |
Statements
UCB revisited: improved regret bounds for the stochastic multi-armed bandit problem (English)
0 references
19 December 2011
0 references
A modification of the UCB algorithm developed for the stochastic multi-armed bandit problem by \textit{P. Auer, N. Cesa-Bianchi} and \textit{P. Fischer} [Mach. Learn. 47, No. 2--3, 235--256 (2002; Zbl 1012.68093)] is presented, which leads to an improvement of the regret bounds of the original UCB algorithm. Both cases of known as well as unknown horizon T are considered.
0 references
stochastic multi-armed bandit problem
0 references
UCB algorithm
0 references
regret bounds
0 references
0 references
