ConceptAI & ML
Vanishing gradients in tanh
Outputs near -1 or 1 in a tanh activation function result in local gradients close to zero, effectively stopping backpropagation.
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
4The lectures that define it or assume it, and the concepts next to it.