Optimal Policy

No Regrets - What Happens to AI Beyond Generative? - Computerphile

Computerphile

The policy that yields the highest expected reward for a given task in reinforcement learning.