DL Notes: Gradient Descent

How Neural Networks “Learn”

Artificial Neural Networks (ANNs) are universal function approximators. They can approximate any complex function if provided with enough data, have a proper architecture, and are trained for enough time.

But what does “training” a network even mean?

In a previous post about the feedforward process, I mentioned that training a network means adjusting the value of its weights, to obtain a better fit of the function we are …

data science deep learning gradient-descent machine learning optimization