Fast automatic step size selection for zeroth-order nonconvex stochastic optimization

作者：

Highlights：

• We propose an automatic step size for zeroth-order nonconvex stochastic optimization.

• We provide a rigorous theoretical analysis of the proposed algorithm.

• Extensive numerical experiments verify the efficacy of the proposed algorithm.

摘要

•We propose an automatic step size for zeroth-order nonconvex stochastic optimization.•We provide a rigorous theoretical analysis of the proposed algorithm.•Extensive numerical experiments verify the efficacy of the proposed algorithm.

论文关键词：Stochastic optimization,Variance reduction,Zeroth-order optimization,Adaptive step size

论文评审过程：Received 24 December 2020, Revised 7 February 2021, Accepted 16 February 2021, Available online 25 February 2021, Version of Record 15 March 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.114749