Coursebooks 2017-2018

• Directory
• EPFL websites
• Place
• Courses
• Publications
• News
• EPFL units

Search

 

Advanced machine learning

MICRO-570

Lecturer(s) :

Billard Aude

Language:

English

Summary

This course will present some of the core advanced methods in the field for structure discovery, classification and non-linear regression. This is an advanced class in Machine Learning; hence, students are expected to have some background in the field.

Content

The class will be accompanied by practical session on computer, using the mldemos software (http://mldemos.epfl.ch) that encompasses more than 30 state of the art algorithms.

• Introduction to the major mathematical principles of Machine Learning
• Structure Discovery: spectral and kernel methods, kernel PCA.CCA, X-means
• Advanced Nonlinear Regression Methods
• Stochastic Modeling: Particle Filters, Reinforcement Learning and Gradient Methods

Keywords

Machine learning, statistics

Learning Prerequisites

Required courses

Probability & Statistics, Linear Algebra

Recommended courses

Machine Learning, Pattern Recognition

Important concepts to start the course

Linear Algebra: Eigenvalue and singular value decomposition

Statistics: Definitions of probability density function, marginal, likelihood, covariance, correlation

Optimization: Lagrange multipliers, gradient descent, local and global optima

 

Learning Outcomes

By the end of the course, the student must be able to:

• Choose an appropriate method
• Apply the method properly

Transversal skills

• Use a work methodology appropriate to the task.
• Write a scientific or technical report.

Teaching methods

Ex-cathedra lectures, exercises, computer-based practical sessions

Expected student activities

Each week, students should read the selected chapters of the Lecture Notes prior to class.

Students must attend the computer-based practice session and prepare regular reports that are graded.

Assessment methods

50% personal work during semester, 50% oral exam

Supervision

Office hours	No
Assistants	Yes
Forum	No

Resources

Ressources en bibliothèque

• Machine Learning Technique / Billard

Notes/Handbook

Machine Learning Techniques, available at the Librairie Polytechnique. To be purchased before the class starts.

Websites

• http://lasa.epfl.ch/teaching/lectures/ML_MSc_Advanced/

Moodle Link

• http://moodle.epfl.ch/course/view.php?id=14885#section-0

Prerequisite for

Students must be knowledgeable about machine learning and have taken a course in the area either at EPFL or elsewhere. Relevant courses at EPFL are:

Applied Machine Learning - MICRO-455

Pattern Classification and Machine Learning: CS-433

Data Analysis and Model Classification - EE-516