1. Feature Visualization One way to make progress in rendering deep artificial neural networks intelligible is to figure out what is going on inside their hidden layers. ‘Hidden’ in this context just means located between the input and output layers1. Drawing...
