Invited talk: On the Theory of Policy Gradient Methods: Optimality, Generalization, and Distribution Shift

ICML 2020