Tight last-iterate convergence rates for no-regret
       learning in multi-player games

     Noah Golowich        Sarath Pattathil        Constantinos Daskalakis
      MIT CSAIL          MIT EECS              MIT CSAIL
     nzg@mit.edu        sarathp@mit.edu         costis@csail.mit.edu

                       Abstract
     We study the question of obtaining last-iterate convergence rates for no-regret
     learning algorithms in multi-player games. We show that the optimistic gradient
     (O ...