arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.00224math.OCcs.DC
收藏

分布式随机非凸优化:基于矩的方差缩减

Distributed Stochastic Non-Convex Optimization: Momentum-Based Variance Reduction

Prashant Khanduri, Pranay Sharma, Swatantra Kafle, Saikiran Bulusu, Ketan Rajawat, Pramod K. Varshney

摘要:

在这项工作中,我们提出了一个分布式算法的随机非凸优化。我们考虑一个工作者-服务器架构,其中一组$K$工作者节点(WN)与服务器节点(SN)合作,共同的目标是最小化一个全球性的,潜在的非凸目标函数。目标函数被假定为在每个WN处可用的局部目标函数的总和,其中每个节点仅能够访问其局部目标函数的随机样本。与现有的方法相比,我们采用了基于动量的“单循环”分布式算法,消除了需要计算大批量梯度,以实现方差减少。我们提出了两种算法,一种是“自适应”学习率,另一种是“非自适应”学习率。我们表明,所提出的算法实现了最佳的计算复杂性,同时达到线性加速与WN的数量。具体地,算法在$T$迭代中利用$\mathbb{E}\| \nabla f(x_a) \| \leq \tilde{O}(K^{-1/3}T^{-1/2} + K^{-1/3}T^{-1/3})$达到$ε$ -稳定点$x_a$,从而需要在每个WN处进行$\tilde{O}(K^{-1} ε^{-3})$梯度计算。此外,我们的方法不假设WN之间的数据分布相同,使得该方法对于联邦学习应用程序来说足够通用。

英文摘要:

In this work, we propose a distributed algorithm for stochastic non-convex optimization. We consider a worker-server architecture where a set of $K$ worker nodes (WNs) in collaboration with a server node (SN) jointly aim to minimize a global, potentially non-convex objective function. The objective function is assumed to be the sum of local objective functions available at each WN, with each node having access to only the stochastic samples of its local objective function. In contrast to the existing approaches, we employ a momentum based "single loop" distributed algorithm which eliminates the need of computing large batch size gradients to achieve variance reduction. We propose two algorithms one with "adaptive" and the other with "non-adaptive" learning rates. We show that the proposed algorithms achieve the optimal computational complexity while attaining linear speedup with the number of WNs. Specifically, the algorithms reach an $ε$-stationary point $x_a$ with $\mathbb{E}\| \nabla f(x_a) \| \leq \tilde{O}(K^{-1/3}T^{-1/2} + K^{-1/3}T^{-1/3})$ in $T$ iterations, thereby requiring $\tilde{O}(K^{-1} ε^{-3})$ gradient computations at each WN. Moreover, our approach does not assume identical data distributions across WNs making the approach general enough for federated learning applications.

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。

↑