In this article of the NNFwP series we'll do the “student-teacher” experiment with two neural networks, where one network will learn directly from the other.
In the fifth article of this short series we will be handling some subtleties that we overlooked in our experiment to classify handwritten digits from the MNIST dataset.
In the fourth article of this short series we will apply our neural network framework to recognise handwritten digits.
The third article of this short series concerns itself with the implementation of the backpropagation algorithm, the usual choice of algorithm used to enable a neural network to learn.
In the second article of this short series we will create a class for a generic neural network and we will also see how to assess the quality of the output of a network, essentially preparing ourselves to implement the backpropagation algorithm.
This is the first article in a series to implement a neural network from scratch. We will set things up in terms of software to install, knowledge we need, and some code to serve as backbone for the remainder of the series.
