Machine Learning CS-433

This course is offered jointly by the TML and MLO groups. Previous year’s website: ML 2020.
See here for the ML4Science projects.

Contact us: Use the discussion forum, or some of the contact details below:

Instructor Nicolas Flammarion Instructor Martin Jaggi
Office INJ 336 Office INJ 341
Email [email protected] Email [email protected]
Office Hours By appointment Office Hours By appointment
Teaching Assistants
  • Maksym Andriushchenko
  • El Mahdi Chayti
  • Lie He
  • Anastasiia Koloskova
  • Tao Lin
  • Amirkeivan Mohtashami
  • Ehsan Pajouheshgar
  • Scott Pesme
  • Thijs Vogels
  • Tianzong Zhang
Student Assistants
  • Raphaël Attias
  • Axel Dinh Van Chi
  • Karim Hadidane
  • Ella Rajaonson
  • Ekrem Yilmazer
  • Yingxue Yu
  • Oguz Yuskel
Lectures Tuesday 17:15 – 19:00 in Rolex Learning Center
  Thursday 16:15 – 18:00 in SG1
Exercises Thursday 14:15 – 16:00

Rooms: INF119INF2INJ218INM202INR219 or zoom (assignment link on moodle)

Language:   English
Credits :   7 ECTS

For a summary of the logistics of this course, see the course info sheet here (PDF).
(and also here is a link to official coursebook information).

Special Announcements

  • Exam date: 20.01.2022 from 08h15 to 11h15 in swisstech
  • The links for the exercises signup and the discussion forum password are on moodle. All other materials are here on this page and github.
  • Projects: There will be two group projects during the course.
    • Project 1 counts 10% and is due Nov 1st.
    • Project 2 counts 30% and is due Dec 23th.
  • The videos of the lectures for each week, new exercise sheet, as well as the solutions for the previous week will be made available each Tuesday. Labs and projects will be in Python. See Lab 1 to get started.
  • Code Repository for Labs, Projects, Lecture notes: github.com/epfml/ML_course
  • the exam is closed book but you are allowed one crib sheet (A4 size paper, both sides can be used), either handwritten or 11 point minimum font size; bring a pen and white eraser; you find the exams from the past years with solutions here:

Detailed Schedule

Annotated lecture notes from each class are made available on github here, and videos here on youtube.

Date Topics Covered Lectures Slides Exercises Projects
21/9 Introduction, Linear Regression 01a,01b
01c,01d
     
23/9 Cost functions     Lab 1  
28/9 Optimization 02a      
30/9 Optimization     Lab 2 Project 1 start
05/10 Least Squares, Overfitting 03a,03b      
07/10 Max Likelihood, Ridge Regression, Lasso 03c,03d   Lab 3  
12/10 Generalization, Model Selection, and Validation 04a 04a    
14/10 Bias-Variance decomposition 04b 04b Lab 4  
19/10 Classification 05a 05a    
21/10 Logistic Regression 05b 05b Lab 5  
26/10 Generalized Linear Models 06a 06a    
28/10 K-Nearest Neighbor 06b 06b Lab 6  
02/11 Support Vector Machines 07a 07a   Proj. 1 due 1.11.
04/11 Kernel Regression 07b 07b Lab 7  
09/11 Neural Networks – Basics, Representation Power 08a,08b 08ab   Project 2 start
11/11 Neural Networks – Backpropagation, Activation Functions 08c,08d 08cd Lab 8  
16/11 Neural Networks – CNN, Regularization, Data Augmentation, Dropout 09a,09b 09ab    
18/11 Adversarial ML 09c 09c Lab 9  
23/11 Ethics and Fairness in ML   10a    
25/11 Unsupervised Learning, K-Means 10b,10c   Lab 10  
30/11 Gaussian Mixture Models 11a      
02/12 EM algorithm 11b   Lab 11  
07/12 Generative adversarial networks 12a      
09/12 SVD and PCA 12b   Lab 12 & Project Q&A  
14/12 Matrix Factorizations 13a      
16/12 Text Representation Learning 13b   Lab 13  
21/12 Guest Lecture        
23/12 Projects       Proj. 2 due 23.12.

Textbooks

(not mandatory)

Gilbert Strang, Linear Algebra and Learning from Data
Christopher Bishop, Pattern Recognition and Machine Learning
Shai Shalev-Shwartz, Shai Ben-David, Understanding Machine Learning
Michael Nielsen, Neural Networks and Deep Learning

Projects & ML4Science

Final projects last year are done either in ML4Science in collaboration with any lab of EPFL, UniL or other academic institution, or the Reproducibility Challenge for ML papers, or one of the predefined ML challenges.

All info about the interdisciplinary ML4Science projects is available on the separate page here.