ConceptAI & ML
Polyak averaging
Keeping a moving average of the parameter vector during training and using it at test time as a smooth ensemble.
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
2Each in their own terms, from the easiest level.
Sergey Levine
An update rule for target network parameters that smoothly interpolates between old and new parameters at every step.
Connections
5The lectures that define it or assume it, and the concepts next to it.