BriefGPT.xyz
Jul, 2019
非凸环境下的分布式学习 -- 第二部分：从鞍点中多项式逃脱
Distributed Learning in Non-Convex Environments -- Part II: Polynomial Escape from Saddle-Points
HTML
PDF
Stefan Vlaski, Ali H. Sayed
TL;DR
分布式学习算法中应用了扩散策略，能够在限制条件较少的情况下从严格的鞍点中逃脱并返回次二级稳定点.
Abstract
The
diffusion strategy
for
distributed learning
from streaming data employs local
stochastic gradient updates
along with exchange of itera
→