EE-613 / 4 credits

Teacher(s): Calinon Sylvain, Canévet Olivier, Odobez Jean-Marc, Villamizar Michael

Language: English

Remark: Next time: Fall 2025


Frequency

Every 2 years

Summary

The objective of this course is to give an overview of machine learning techniques used for real-world applications, and to teach how to implement and use them in practice. Laboratories will be done in python using jupyter notebooks.

Content

Fundamentals

  • Recalls in probability and information theory
  • Notion of learning, cross validation and performance evaluation
  • Optimization (gradient, Newton method, stochastic gradient, etc.)

 

Regression

  • Least squares (Tikhonov regularization)
  • Weighted least squares
  • Iteratively reweighted least squares (IRLS)
  • Tensor regression
  • Gaussian mixture regression (GMR)
  • Gaussian process regression (GPR)

 

Generative models

  • Bayesian networks,
  • Directed / non-directed models, conditional independence, Naive Bayes
  • k-Means, Gaussian mixture model (GMM), Expectation-Maximization (EM)
  • PCA and probabilistic PCA
  • Hidden Markov model (HMM)

 

Discriminative models

  • Logistic regression
  • Decision trees, random forest
  • Support vector machine (SVM) and Kernelization (PCA, etc.)

 

Deep learning

  • Perceptron, Multi-Layer-Perceptron (MLP)
  • Convolutional neural network (CNNs)

 

Meta-algorithms

  • Ensemble methods
  • Bagging
  • Boosting

Learning Prerequisites

Required courses

  • Undergraduate knowledge of probabilities, linear algebra, and statistics
  • Python programming

Learning Outcomes

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

  • Understand the core principles of machine learning and of the different concepts and algorithms behind the different learning methodologies.
  • Select appropriately in practice standard learning-based inference techniques for regression, classification and density modeling, including understanding the impact of different parameters

Assessment methods

Multiple.

Resources

Bibliography

Pattern Recognition and Machine Learning,

C. Bishop,

Springer, 2008

Ressources en bibliothèque

Moodle Link

In the programs

  • Number of places: 50
  • Exam form: Multiple (session free)
  • Subject examined: Machine Learning for Engineers
  • Lecture: 28 Hour(s)
  • Practical work: 28 Hour(s)
  • Type: optional
  • Number of places: 50
  • Exam form: Multiple (session free)
  • Subject examined: Machine Learning for Engineers
  • Lecture: 28 Hour(s)
  • Practical work: 28 Hour(s)
  • Type: optional
  • Number of places: 50
  • Exam form: Multiple (session free)
  • Subject examined: Machine Learning for Engineers
  • Lecture: 28 Hour(s)
  • Practical work: 28 Hour(s)
  • Type: optional

Reference week

Related courses

Results from graphsearch.epfl.ch.