Graduate School and Research Center in Digital Sciences

Intelligent systems

T Technical Teaching


  • The objective of this course is to give student a solid background on techniques for classification and learning. The relationship with intelligence is that those techniques are often useful to build effective models in situations where no optimal solution in known, for example fraud detection in credit card usage. The resulting systems can be considered as having some kind of intelligent behaviour.


  • Search Techniques: Artificial intelligence. Winston, Patrick Henry, Addison-Wesley 1992, ISBN : 0201533774
  • Neural Networks:
  • Perceptrons : an introduction to computational geometry. Minsky, Marvin, Papert, Seymour MIT Press 1988, ISBN : 0262631113
  • Neural networks : a systematic introduction. Rojas, Raul Springer-Verlag 1996, ISBN : 3540605053
  • Genetic Algorithms: An Introduction to genetic algorithms Mitchell, Melanie MIT Press 1996, ISBN : 0262133164


  • C / C++ using Visual Studio for programming tasks during the lab sessions


  • This course will cover some basic and advanced techniques for classification and optimisation : Neural Networks, Genetic Algorithms, Simulated Annealing, Decision Trees and Bayesian Networks. Applications such as Intelligent Agents and Data mining will illustrate the practical usage.
  • A large part of the course will be devoted to the study of Neural Networks, which are one of the most popular methods, and as proved to be quite effective in many situations. We will study the basic perceptron, multi-layer perceptrons, the back-propagation algorithm.
  • We will also take a look to other types of networks such as the Hopfield networks. We will explore advanced techniques for optimisation, such as Genetic Algorithms and Simulated Annealing.
  • We will also study the construction of classification models such as Decision Trees and Bayesian Networks. Throughout the course, we will illustrate the usage of those techniques in applications such as Intelligent Agent, Knowledge Discovery and Data Mining
Nb hours: 42.00
Nb hours per week: 3.00
Control form: examen écrit