ConceptAI & ML
Block coordinate descent for hyperparameters
A sequential tuning strategy where you find a good learning rate first and then go back to try different model sizes to reduce exponential search complexity.
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
2The lectures that define it or assume it, and the concepts next to it.