Presentation
Machine learning is the part of artificial intelligence that relies on mathematical models to enable computers to learn and perform tasks from data. The main objective of this school is to present the general concepts of machine learning (Machine Learning -ML- and Deep Learning -DL-) and to define its fields and conditions of application through concrete cases. This school is funded by the CNRS AISSAI (Artificial Intelligence for Science and Science for Artificial Intelligence) project.
Objectives
At the end of the course, participants will be able to :
- Identify the nature of a machine learning problem: supervised / unsupervised, classification / regression
- Understand the mathematical concepts of classical ML and DL methods
- Implement popular ML methods (SVM, decision trees, etc.)
- Implement a simple neural network architecture (Multi-Layer Perceptron and Convolution Network)
- Learn about the main DL algorithms
- evaluate the performance of these methods using a range of metrics
- Interpret algorithm results and identify their limitations
- Use Sklearn, Keras / Tensor Flow tools.
The week will be organized in the form of alternating theoretical lectures and computer-based practical exercises.
Audiences
Cette école est ouverte aux chercheurs (y compris doctorants et post-doctorants) et aux ingénieurs de tous instituts.
Langage
Classes will be taught in French.
Prerequisite
Good knowledge of Python and NumPy; basic knowledge of linear algebra/statistics/probability.
Equipement
Trainees are asked to bring their own computer, on which the necessary
on which the free software required for the training installed beforehand. Instructions will be given before the beginning of the course.
Participation is limited to 50 participants. Consequently, registration does not constitute acceptance, and a selection may be made if necessary.
Location
Training will take place on the Orsay-Vallée campus, at IJCLab in building 100:
Monday: Joliot Curie auditorium (basement)
Tuesday to Friday: boards room (ground floor)
