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 2106.09481math.OCcs.DScs.LG
收藏

随机偏置-约化梯度法

Stochastic Bias-Reduced Gradient Methods

Hilal Asi, Yair Carmon, Arun Jambulapati, Yujia Jin, Aaron Sidford

摘要:

我们开发了一个新的原始随机最优化:一个低偏差,低成本的最小化$x_\star$的任何Lipschitz强凸函数的估计。特别是,我们使用多层次的蒙特卡罗方法,由于Blanchet和Glynn把任何最佳的随机梯度法的估计$x_\star$偏$δ$,方差$O(\log(1/δ))$,和预期的采样成本的$O(\log(1/δ))$随机梯度评估。作为一个直接的后果,我们获得廉价的和几乎无偏梯度估计的Moreau-Yoshida包络的任何Lipschitz凸函数,使我们能够执行无量纲随机平滑。我们通过四个应用程序证明了我们的估计的潜力。首先,我们开发了一种方法,用于最大限度地减少$N$功能,改善最近的结果和匹配的下界的对数因子。第二和第三,我们使用简单的算法和透明的分析来恢复投影效率和梯度效率优化的最新速率。最后,我们表明,我们的估计的改进版本将产生一个接近线性时间,最优效用,微分私人非光滑随机最优化方法。

英文摘要:

We develop a new primitive for stochastic optimization: a low-bias, low-cost estimator of the minimizer $x_\star$ of any Lipschitz strongly-convex function. In particular, we use a multilevel Monte-Carlo approach due to Blanchet and Glynn to turn any optimal stochastic gradient method into an estimator of $x_\star$ with bias $δ$, variance $O(\log(1/δ))$, and an expected sampling cost of $O(\log(1/δ))$ stochastic gradient evaluations. As an immediate consequence, we obtain cheap and nearly unbiased gradient estimators for the Moreau-Yoshida envelope of any Lipschitz convex function, allowing us to perform dimension-free randomized smoothing. We demonstrate the potential of our estimator through four applications. First, we develop a method for minimizing the maximum of $N$ functions, improving on recent results and matching a lower bound up to logarithmic factors. Second and third, we recover state-of-the-art rates for projection-efficient and gradient-efficient optimization using simple algorithms with a transparent analysis. Finally, we show that an improved version of our estimator would yield a nearly linear-time, optimal-utility, differentially-private non-smooth stochastic optimization method.

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。

↑