Instructions

Exercises for practical sessions and project work of Artificial Intelligence and Deep Learning

Academic year: 2026-2027¶

Exercises¶

• Practicals 1 (recap)

• Practicals 2

• Practicals 3

• Mini project

• Project

Objectives¶

The objective of the practical work is to prepare the project by practising the different aspects of artificial intelligence and deep learning. Following the completion of all the exercises, you will be able to

• implement the following topics:

  • symbolic artificial intelligence

  • artificial neural networks

  • deep learning

• implement a perceptron and a multilayer perceptron (forward propagation, backpropagation) in NumPy

• create neural network models using different structures and activation functions

• configure and optimize neural networks

• test existing models of recurrent neural networks such as LSTM and convolutional neural networks

• write and execute simple programs in the Prolog programming language

Instructions¶

1. All practical work is based on the aspects you have already seen during your course. Make good use of your course materials.

2. Work in pairs.

3. It is mandatory to cite all sources (e.g. internet, groups, AI assistants).

4. The sessions are supervised by 2 teachers.

Evaluation¶

1. Only the project is graded: it accounts for 100% of the final grade.

2. The three practicals are not graded and are not to be submitted. They let you apply the course material and prepare for the project.

3. The grading criteria of the project are described in the project description.

4. The project deadline is given on e-campus.

5. Each practical question has a difficulty level

   • ★: Easy

   • ★★: Average difficulty

   • ★★★: Difficult

Project submission¶

Submission is online (e-campus) and concerns the project only.

Your submission folder should contain the following files:

a. README: the project report (see [the project description](Project/project.md))
b. CONTRIBUTORS: names and first names of contributors
c. src/: the source code (notebooks and scripts) in the src directory

Your submission must be renamed as group_N1_N2, where N1 and N2 are the names (e.g., group_DUPONT_SMITH).

Write README and CONTRIBUTORS in markdown format.

Please do not submit datasets (CSV, JSON, TSV, image or audio files) or trained models: explain instead how to obtain or regenerate them.

Check list¶

Before submitting your project, verify this checklist:

• ✅ The names (first name and last names) of the pair are present in the CONTRIBUTORS file

• ✅ The README file is completely filled.

• ✅ All sources are cited.

• ✅ Your code is well commented.

• ✅ Your code can be executed without any error (and if possible, without any warning).

Code template¶

For reference, you can consult the following repository: https://github.com/johnsamuelwrites/IA-DeepLearning. It contains code templates for the exercises of the three practicals and for the project.

You can view it online using the link above or clone it on your machine using the following commands on the terminal.

          $ git clone https://github.com/johnsamuelwrites/IA-DeepLearning
          $ cd IA-DeepLearning
          $ ls
          $ cat en/README.md

You can run the following command for obtaining the latest modifications on the repository:

          $ git pull