arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1605.08481cs.LG
收藏

开放问题:Best Arm识别:几乎实例最优性和间隙熵猜想

Open Problem: Best Arm Identification: Almost Instance-Wise Optimality and the Gap Entropy Conjecture

Tsinghua University(清华大学)

Lijie Chen, Jian Li

摘要:

最佳手臂识别问题(BEST-1-ARM)是随机多臂强盗中最基本的纯探索问题。这个问题由来已久,在过去十年中引起了极大的关注。然而,我们还没有一个完整的理解的最佳样本复杂性的问题:国家的最先进的算法实现了样本复杂性$O(\sum_{i=2}^{n} Δ_{i}^{-2}(\lnδ^{-1} + \ln\lnΔ_i^{-1}))$($Δ_{i}$是最大的平均值和$i^{th}$平均值之间的差异),而最知名的下限是$Ω(\sum_{i=2}^{n} Δ_{i}^{-2}\lnδ^{-1})$的一般情况和$Ω(Δ^{-2} \ln\ln Δ^{-1})$的双臂实例。我们建议研究最佳1臂问题的实例最优性。以往的工作已经证明,这是不可能有一个实例优化算法的2臂问题。然而,我们猜想,模的加法项$Ω(Δ_2^{-2} \ln\ln Δ_2^{-1})$(这是一个上限和最坏情况下的2-arm问题的下限),有一个实例优化算法的最佳-1-ARM。此外,我们引入了一个新的量,称为间隙熵的最佳-arm问题的实例,并推测它是实例的下限。因此,解决这个猜想将为这个古老而基本的问题提供最终答案。

英文摘要:

The best arm identification problem (BEST-1-ARM) is the most basic pure exploration problem in stochastic multi-armed bandits. The problem has a long history and attracted significant attention for the last decade. However, we do not yet have a complete understanding of the optimal sample complexity of the problem: The state-of-the-art algorithms achieve a sample complexity of $O(\sum_{i=2}^{n} Δ_{i}^{-2}(\lnδ^{-1} + \ln\lnΔ_i^{-1}))$ ($Δ_{i}$ is the difference between the largest mean and the $i^{th}$ mean), while the best known lower bound is $Ω(\sum_{i=2}^{n} Δ_{i}^{-2}\lnδ^{-1})$ for general instances and $Ω(Δ^{-2} \ln\ln Δ^{-1})$ for the two-arm instances. We propose to study the instance-wise optimality for the BEST-1-ARM problem. Previous work has proved that it is impossible to have an instance optimal algorithm for the 2-arm problem. However, we conjecture that modulo the additive term $Ω(Δ_2^{-2} \ln\ln Δ_2^{-1})$ (which is an upper bound and worst case lower bound for the 2-arm problem), there is an instance optimal algorithm for BEST-1-ARM. Moreover, we introduce a new quantity, called the gap entropy for a best-arm problem instance, and conjecture that it is the instance-wise lower bound. Hence, resolving this conjecture would provide a final answer to the old and basic problem.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑