Introduction aux réseaux de neurones profonds et à convolution | Introduction to Deep and Convolutional Neural Networks

Deep neural networks (DNN) simulate the way that neurons work in brains and these types of models are surprisingly good at pattern detection which makes them ideal for machine learning. In particular convolutional neural networks (CNN) excel at image recognition.

This module consists of three notebooks. "1 - DNN" goes into the core principles of DNN and how to build them from first principles. In the second notebook, "2 - DNN_PyTorch", we will use a dedicated Python library to do the heavy lifting and build more complex neural networks. Finally, in the last notebook, "3 - CNN_Concepts", we will look at image recognition and specifically recognition of hand-written digits.