Exploding Gradient
Defined in 2 lectures. Each definition is in that lecture's terms.
Definitions are written by AI from each lecture’s transcript. The time opens the lecture where the lecturer explains it.
- A scenario where the gradient goes to infinity as the minimizer is approached, causing unstable updates and overshooting.
- A problem where weight matrices are very large, causing gradients to grow exponentially during backpropagation and making network training unstable.