The concept is not easy but can be visualised by showing back propagation.

What is gradient descent? I promise, no calculus.

Gradient descent is the algorithm to nudge the weights in a neural network into the right direction. The concept is not easy but can be visualised by showing the back propagation of a neural network with one and three inputs.

Want to see more?

Gradient descent is a big part of my 'Neural Network from scratch in Python'-course. If you like my style of teaching, you can support me by enrolling in the online course on Udemy.

Written by Loek van den Ouweland on 2021-10-16.
Questions regarding this artice? You can send them to the address below.