ConceptAI & ML
Offline RL regularizer vs supervised learning regularizer
Unlike supervised learning regularizers which only penalize feature norms, offline RL regularizers include an additional dot-product term between current and next-state features due to target dependence on model predictions.
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.