Global optimality of softmax policy gradient with single hidden layer neural networks in the mean-field regime.

Andrea Agazzi, Jianfeng Lu

Browse the full ICLR paper archive.