arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.07387cs.LG
收藏

伪一维Bandit凸优化的最优后悔算法

Optimal Regret Algorithm for Pseudo-1d Bandit Convex Optimization

Microsoft Research(微软研究院);Google Research India(谷歌印度研究院)

Aadirupa Saha, Nagarajan Natarajan, Praneeth Netrapalli, Prateek Jain

摘要:

我们研究在线学习的强盗反馈(即学习者只能访问零阶预言)的成本/回报函数$\f_t$承认一个“伪1d”的结构,即$\f_t(\w) = \loss_t(\pred_t(\w))$的输出$\pred_t$是一维的。在每一轮,学习者观察上下文$\x_t$,为一些$\w_t \in \mathbb{R}^d$播放预测$\pred_t(\w_t; \x_t)$(例如$\pred_t(\cdot)=\langle \x_t, \cdot\rangle$),并观察损失$\loss_t(\pred_t(\w_t))$,其中$\loss_t$是凸Lipschitz连续函数。我们的目标是最小化标准的后悔度量。这种伪一维强盗凸优化问题(\SBCO)经常出现在大型系统的在线决策或参数调整等领域。对于这个问题,我们首先显示了一个下界的$\min(\sqrt{dT}, T^{3/4})$的遗憾的任何算法,其中$T$是轮数。我们提出了一个新的算法\sbcalg,结合随机在线梯度下降与核化指数权重的方法,有效地利用伪一维结构,保证{ \em最优}遗憾界上述,额外的对数因子。相比之下,应用最先进的在线凸优化方法会导致$\tilde{O}\left(\min\left(d^{9.5}\sqrt{T},\sqrt{d}T^{3/4}\right)\right)$遗憾,这在$d$中明显是次优的。

英文摘要:

We study online learning with bandit feedback (i.e. learner has access to only zeroth-order oracle) where cost/reward functions $\f_t$ admit a "pseudo-1d" structure, i.e. $\f_t(\w) = \loss_t(\pred_t(\w))$ where the output of $\pred_t$ is one-dimensional. At each round, the learner observes context $\x_t$, plays prediction $\pred_t(\w_t; \x_t)$ (e.g. $\pred_t(\cdot)=\langle \x_t, \cdot\rangle$) for some $\w_t \in \mathbb{R}^d$ and observes loss $\loss_t(\pred_t(\w_t))$ where $\loss_t$ is a convex Lipschitz-continuous function. The goal is to minimize the standard regret metric. This pseudo-1d bandit convex optimization problem (\SBCO) arises frequently in domains such as online decision-making or parameter-tuning in large systems. For this problem, we first show a lower bound of $\min(\sqrt{dT}, T^{3/4})$ for the regret of any algorithm, where $T$ is the number of rounds. We propose a new algorithm \sbcalg that combines randomized online gradient descent with a kernelized exponential weights method to exploit the pseudo-1d structure effectively, guaranteeing the {\em optimal} regret bound mentioned above, up to additional logarithmic factors. In contrast, applying state-of-the-art online convex optimization methods leads to $\tilde{O}\left(\min\left(d^{9.5}\sqrt{T},\sqrt{d}T^{3/4}\right)\right)$ regret, that is significantly suboptimal in $d$.

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。

↑