An experimental design perspective on model-based reinforcement learning

Reinforcement learning (RL) has achieved astonishing successes in domains where the environment is easy to simulate. For example, in games like Go or those in the Atari library, agents can play millions of games in the course of days to explore the environment and find superhuman policies. However, transfer of these advances to broader real-world applications is challenging because the cost of exploration in many important domains is high.

deep dive design experimental learning perspective reinforcement reinforcement learning research