[R] Anyway to measure the importance of each neural network layer? • r/MachineLearning

Given a neural network composed of N layers, usually some layers are more "important" than other layers (eg. Is there any theory that can be used to measure the effectiveness of each neural network layer?