The invention of convolutional neural networks (CNNs) applied to vision represents by far one of the most innovative achievements in the history of applied mathematics. With its multiple layers (visible and hidden), CNN has brought artificial intelligence from machine learning to deep learning.
A CNN relies on two basic tools of linear algebra: kernels and applying them to convolutions as described in this chapter. These tools have been used in applied mathematics for decades.
However, it took the incredible imagination of Yan LeCunn, Yoshua Bengio, and others—who built a mathematical model of several layers—to solve real-life problems with CNNs.
This chapter describes the marvels of CNNs, one of the pillars of Artificial Neural Networks (AANs). A CNN will be built from scratch, trained, and saved. The classification model described...
