BriefGPT.xyz
Oct, 2016
基于方差减少的随机优化算法在具有有限和结构的无限数据集上的应用
Stochastic Optimization with Variance Reduction for Infinite Datasets with Finite-Sum Structure
HTML
PDF
Alberto Bietti, Julien Mairal
TL;DR
本文提出了针对复合目标强凸的情况下，带有方差约束的随机梯度下降法，其收敛速度优于传统的随机梯度下降法，同时常数因子也更小，只与输入数据的方差有关。
Abstract
stochastic optimization
algorithms with
variance reduction
have proven successful for minimizing large finite sums of functions. However, in the context of empirical risk minimization, it is often helpful to augm
→