Practical session on
simple "shallow" neural networks
(Multi-Layer Perceptrons, MLP)

Short introduction to "Jupyter notebooks"

Jupyter Notebook is an interactive environment for programming and executing programs, mainly in Python language. In a notebook (file with .ipynb extension) can be combined text, images and links (as in an HTML file), but also Python code that can be modified and executed DIRECTLY in the notebook page. The results of code execution (printed text, plots, images, etc...) are automatically displayed just below the executed code.
This environment, as well as the SciKit-learn library are now very commonly used in the Machine-Learning (ML) community.

Preliminary configuration

IF YOU WORK ON COMPUTERS IN ONE OF THE PRACTICAL SESSION ROOMS OF MINES ParisTech (L.117, L.119, L.120, L.022, etc), NOTHING NEEDS TO BE INSTALLED, BUT YOU NEED TO BOOT IN **LINUX**.
Just launch a Terminal window, and first ADAPT THE VALUE OF the "PATH" variable:
  export PATH=/opt/anaconda3/bin:$PATH
Then, start jupyter-notebook in the terminal, by just typing:
  jupyter-notebook
 This shall open *in your web navigator* a kind of "explorer" allowing you to find, open, and then execute any notebook (files with .ipynb extension) present on your computer.

If you work somewhere else, or on your own laptop, then you must first make sure that Python, SciKit-Learn, and jupyter-notebook are properly installed.
The easiest way to install all you need on ANY computer (e.g. your own laptop) with ANY OS (Linux, Windows or MacOS) is to INSTALL  Anaconda, which is a large "all in one" package (Jupyter, python, and common libraries) existing for ALL common Operating Systems (Windows, MacOS, and Linux). Once this software is installed on your computer, you shall only need to start Jupyter-notebook, which will open in your web navigator a kind of "explorer" allowing you to find, open, and then execute any notebook (files with .ipynb extension)present on your computer.

Note that it is also possible to run the notebook distantly by cloud computing on GoogleCOLAB (see instructions below).

Starting the practical session

If you work LOCALLY, first download somewhere on your computer the notebook of this practical session: mlp-notebook.ipynb 
Then, start jupyter-notebook, search, find and open the notebook you have just downloaded.

If you want to run the notebook distantly on Google COLAB, you only need to open the link below which points to the notebook verion stored on GoogleDrive : https://drive.google.com/file/d/1VlfC8fQ0gBdIvkylIlsFYP1wt7MxzrK_/view?usp=sharing
and then choose "Open with Google Colaboratory" (Note that you will need to connect to your Google account).

Assignments of the practical session

They are described directly in the notebook, so you shall read them after opening it with Jupyter. However, you can also view them before in non-interactive HTML format if you wish.