Member-only story
A Guide to Artificial Neural Network Visualizations
I originally wanted to make this a post about making Real Time Neural Network Visualizations which is what I need for my current projects but I think the topic is interesting enough to have an overview on the existing Artificial Neural Network Visualizations out there ( regardless of the real time element ).
WHY ? Neural Networks can be complex things, so having a visual aid can help understand and develop new ones, it worked for Turing, it should work for us.
— — — — — — — ⭐️SUBSCRIBE TO MEDIUM...( For good luck ! ) ⭐️— — — — — — —
Model views
Let’s start with model views in pseudo graph notation and the original Neural Network, The perceptron :
This will either be a simple and trivial illustration for those into ANNs or scary and complex if you are just getting started, I like to think it condenses ~ 2.5 million years of human Science and Technology, so there’s that.
For those in the second camp I tried explaining them from the ground up in simple terms:
You can go surprisingly far with this simple graphical representation as you can add layers, mechanics, pre and post processing, etc,etc.
There’s a few new things here, for one this is a Deep Neural Network as it has more than one intermediate layer ( it has 2 ), these layers are usually called hidden layers, there’s also a backward connection between layers (in black ) to indicate this is a recurrent deep neural network , here’s a list of a few more to give you an idea of the variety:
