Adaptive Experience Selection for Policy Gradient

Saad Mohamad, Giovanni Montana

Policy gradient reinforcement learning (RL) algorithms have achieved impressive performance in challenging learning tasks such as continuous control, but suffer from high sample co…