A Derivation of Feedforward Neural Network Gradients Using Fréchet Calculus

Thomas Hamm

We present a derivation of the gradients of feedforward neural networks using Fréchet calculus which is arguably more compact than the ones usually presented in the literature. We…