Conditional Accelerated Lazy Stochastic Gradient Descent

Conditional Accelerated Lazy Stochastic Gradient Descent
复制标题

条件加速惰性随机梯度下降

DOI:
--
复制
发表时间:
2017
期刊:
International Conference on Machine Learning
影响因子:
--
通讯作者:
Daniel Zink
Daniel Zink
中科院分区:
--
文献类型:
--
作者:
Guanghui Lan;S. Pokutta;Yi Zhou;Daniel Zink

文献摘要

被引文献

相似文献

在这项工作中,我们引入了一个条件加速惰性随机梯度下降算法,具有最佳的调用次数和收敛速度$Oleft(frac{1}{varepsilon^2} ight)$改进Hazan和Kale [2012]的无投影、基于在线Frank-Wolfe的随机梯度下降,收敛速度为$Oleft(frac{1}{varepsilon^4} 八)$。
In this work we introduce a conditional accelerated lazy stochastic gradient descent algorithm with optimal number of calls to a stochastic first-order oracle and convergence rate $Oleft(frac{1}{varepsilon^2} ight)$ improving over the projection-free, Online Frank-Wolfe based stochastic gradient descent of Hazan and Kale [2012] with convergence rate $Oleft(frac{1}{varepsilon^4} ight)$.