"Look deep into nature, and then you will understand everything better"
– Albert Einstein
No chapter about the use of neural networks to classify images is allowed to end without touching on CNNs. Despite the fact that scikit-learn does not implement convolutional layers, we can still understand the concept and see how it works.
Let's start with the following 5 x 5 image and see how to apply a convolutional layer to it:
x_example = array(
[[0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 1, 1, 0], [0, 0, 1, 1, 0], [0, 0, 0, 0, 0]]
)
In natural language processing, words usually serve as a middle ground between characters and entire sentences when it comes to feature extraction. In this image, maybe smaller patches serve as better units of information than a separate pixel. The objective of this section is to find ways to represent these small 2 x 2, 3 x 3, or N x N patches...
