BriefGPT.xyz
May, 2016
反演问题中收敛速度与重构精度之间的权衡
Tradeoffs between Convergence Speed and Reconstruction Accuracy in Inverse Problems
HTML
PDF
Raja Giryes, Yonina C. Eldar, Alex M. Bronstein, Guillermo Sapiro
TL;DR
本文旨在探讨如何在迭代算法中通过低维近似来实现更快速的收敛速度，在保证误差率时，本方法可能会牺牲额外的重构误差。
Abstract
Solving
inverse problems
with
iterative algorithms
such as stochastic gradient descent is a popular technique, especially for large data. In applications, due to time constraints, the number of iterations one may
→