Bounded Regret for Finite-Armed Structured Bandits

Bounded Regret for Finite-Armed Structured Bandits
复制标题

DOI:
--
复制
发表时间:
2014-11
期刊:
ArXiv
影响因子:
--
通讯作者:
Tor Lattimore;R. Munos
Tor Lattimore;R. Munos
中科院分区:
其他
文献类型:
--
作者:
Tor Lattimore;R. Munos

文献摘要

被引文献

相似文献

研究了一类新的K-臂强盗问题,其中一个臂的期望收益可能依赖于其他臂的收益。我们提出了一个新的算法,这一般类问题,并表明,在某些情况下,它是可能实现有限的预期累积遗憾。我们还给出了问题相关的累积遗憾的下界,表明至少在特殊情况下,新算法是接近最优的。
We study a new type of K-armed bandit problem where the expected return of one arm may depend on the returns of other arms. We present a new algorithm for this general class of problems and show that under certain circumstances it is possible to achieve finite expected cumulative regret. We also give problem-dependent lower bounds on the cumulative regret showing that at least in special cases the new algorithm is nearly optimal.