Methods of Artificial Intelligence

Abstract

This practical work booklet is part of the course Methods in Artificial Intelligence designed for 4th year Industrial Engineering students at the Higher School of Applied Sciences. Its main objective is to introduce students to the fundamental concepts of artificial intelligence through a practical and progressive approach. The booklet contains five lab sessions aimed at enabling students to master key concepts and develop their own intelligent systems, initially limiting the use of libraries and predefined functions. This approach seeks to deepen understanding of the underlying algorithms and encourage a reflective approach to data processing. The purpose of the first two labs is to implement an intelligent system step by step, without using predefined functions, whether in MATLAB or Python. The first lab guides students through classification using the k-nearest neighbors (k-NN) method, including data cleaning and preparation from a downloaded dataset. The second lab involves creating and using a perceptron, also fully programmed by the students without relying on built-in functions. The third and fourth labs deal with handwritten character recognition using a dataset constructed by the students themselves. The third lab, conducted in Matlab, covers data preparation and the use of a multilayer neural network with the help of MATLAB toolbox and its predefined functions. The fourth lab addresses the same problem in Python, this time utilizing Python libraries such as Scikit-learn and TensorFlow. Finally, the fifth lab focuses on designing a fuzzy system for controlling an inverted pendulum. This is done in MATLAB by first using the dedicated toolbox, then programming the fuzzy system using predefined functions, and finally using the Scikit-Fuzzy library of Python. This booklet aims to provide a balance between theoretical understanding and practical skills, offering students the necessary foundations to design, program, and analyze complete intelligent systems. We hope this work will effectively contribute to the training of future industrial engineers by developing their analytical, programming, and innovation capabilities in the field of artificial intelligence.