Independent Natural Policy Gradient always converges in Markov Potential Games.

Roy Fox, Stephen M. McAleer, Will Overman, Ioannis Panageas

Browse the full AISTATS paper archive.