Last Iterate is Slower than Averaged Iterate in Smooth Convex-Concave Saddle Point Problems

Golowich Noah
Golowich Noah
Pattathil Sarath
Pattathil Sarath

COLT, pp. 1758-1784, 2020.

Cited by: 15|Views24
EI
Weibo:
In this paper we study the smooth convex-concave saddle point problem

Abstract:

In this paper we study the smooth convex-concave saddle point problem. Specifically, we analyze the last iterate convergence properties of the Extragradient (EG) algorithm. It is well known that the ergodic (averaged) iterates of EG converge at a rate of $O(1/T)$ (Nemirovski, 2004). In this paper, we show that the last iterate of EG con...More

Code:

Data:

Your rating :
0

 

Tags
Comments