This repository is use to keep track of my advancement in my 0Reinforcement Learning course. It also contains the basics of Deep Neural Networks such as the use of Convolutionnal Neural Network and Artificial Neural Network using Tensorflow 2.0. Feel free to use the code as you wish :D
The ANN section contains the tools to do data science: How to load, sparse and segment database, create neurons, choose activation function, and process the back propagation. The example are done using a large database of houses to sell in order to predict price according to multiple features such as the location, the surface... This section also covers the basic tool of Machine learning for linear regression using SciKit Library.
The CNN section cover the use of CNN to sparse pictures and extract features that helps the Neural Network to understand parterns. This it quite intersting if you are working with images.
Classical Q learning section is the example used to apply and understand Q learning. It helps to define the problematic and choose the best option according to. Reinforcement Learning is a tool and does not suits every need. For example simple problems can be easly acheived using classic approach of Q learning if the Q space is limited to a reasonable abmount values. In this example we solved an easy example of Gym games using this technique.
Not done yet