摘要
针对信号和图像处理、统计学和机器学习中的正则化风险最小化问题,把快速临近随机方差缩减梯度算法(FISTA-Prox-SVRG)和Polyak步长方法相结合,得到一种新的快速临近随机方差缩减梯度算法.通过数值实验将新算法FISTA-Prox-SVRG-Polyak与已有的两种算法FISTA-Prox-SVRG-BB、FISTA-Prox-SVRG进行比较,结果表明新算法是可行有效的.
To solve regularized risk minimization problems in signal/image processing,statistics and machine learning,we propose a new fast proximal stochastic gradient method with progressive variance reduction called FISTA-Prox-SVRG-Polyak.The new algorithm combines the fast proximal stochastic gradient method with progressive variance reduction with a Polyak step size method.Numerical experiments compared FISTA-Prox-SVRG-Polyak and FISTA-Prox-SVRG algorithms,and the results show that the proposed algorithm is feasible and effective.
作者
史鲁玉
王福胜
SHI Luyu;WANG Fusheng(School of Mathematics and Statistics,Taiyuan Normal University,Shanxi Jinzhong 030619,China)
出处
《太原师范学院学报(自然科学版)》
2023年第3期13-18,共6页
Journal of Taiyuan Normal University:Natural Science Edition
基金
山西省基础研究计划(自由探索类)面上项目(202103021224303).