ConceptAI & ML
Model-based offline policy optimization
MOPO is an offline policy optimization method that modifies the reward function by subtracting an uncertainty penalty to prevent the policy from exploiting model errors.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Lecturers
1Connections
3The lectures that define it or assume it, and the concepts next to it.