Machine learning

Course description
The goal of this course is to present the general principles and methodology of statistical machine-learning algorithms, and a typology of learning tasks (supervised classification or regression, unsupervised clustering etc.). At the same time, students will be provided with quite a wide range of the most commonly used ‘standard’ machine-learning algorithms will be provided to the students (Support Vector Machines, Decision Trees and Random Forests, boosting, neural networks, k-Means, Kohonen Self-Organizing Maps, etc.), as well as the specifics of the Deep-Learning approach, and details of main Deep-Learning algorithms (Convolutional Neural Networks, etc.) The course also includes a significant proportion of practical work on the computer, to provide “hands-on” experience and understanding of main parameters of algorithms.

Objectives

• Introduction, principles and methodology of Machine-Learning
• Support Vector Machines and kernel methods
• Decision Trees and Random Forests, boosting
• Multi-layer neural networks
• Convolutional Neural Network and Deep-Learning
• Unsupervised learning and clustering (k-means, Kohonen SOM, etc.)
• Genetic Algorithms and other meta-heuristics

Academic Instructor
Fabien Moutarde|MINES ParisTech