Global Convergence and Variance Reduction for a Class of Nonconvex-Nonconcave Minimax Problems

Nonconvex minimax problems appear frequently in emerging machine learning applications, such as generative adversarial networks and adversarial learning.