Introduction to machine learning (BA4)

Weekly outline

• 

  General

  Welcome to the machine learning class!

  Machine learning and data analysis are becoming increasingly central in many sciences and applications. In this course, fundamental principles and methods of machine learning will be introduced, analyzed and practically implemented using Python.

  
  

  Instructor

  Prof. Pascal Fua

  Computer Vision Laboratory (CVLAB) 

  Office: BC 310

  E-mail: pascal.fua@epfl.ch

  
  

  Teaching Assistants

  • Nicolas Talabot (nicolas.talabot@epfl.ch)
  • Alexandre de Skowronski (alexandre.deskowronski@epfl.ch)
  • Nikita Durasov (nikita.durasov@epfl.ch)
  • Ren Li (ren.li@epfl.ch)
  • Saqib Javed (saqib.javed@epfl.ch)
  • Tianzong Zhang (tianzong.zhang@epfl.ch)

  Student Assistants

  • Christelle Lam
  • Eugène Bergeron
  • Paul Teiletche
  • Aitor Ganuza Izagirre

  
  

  Organization 

  The lectures will be taught live, during the official class hours, i.e., Tuesdays 8:15-10:00 in CM3.
  
  The class has a SWITCHtube channel where we have the links of the 2020/2021 lectures (link to the channel: https://tube.switch.ch/channels/97cd11b0 ). The content of the lectures will be similar, but there may be changes so we recommend for you to attend the lectures in person.
  
  The exercise sessions will be held in-person weekly in CE1100, CE1101, CE1103 on Tuesdays 10:15-12:00 and will consist of implementing the machine learning methods learned in class. The TAs and SAs will be present to answer questions. 
  
  

  CE1100: Students with last names starting with A - F

  CE1101: Students with last names starting with G - Matthey-Doret

  CE1103:  Students with last names starting with Mellouk - Z

  Student Forum

  If you have any questions please post them in the student forum and we will answer you.

  Project

  We will organise a small project in teams of 3, worth 20% of your final grade. There will be two milestones, each worth 10% of your final grade, at the end of week 8 and 14.

  Final Exam

  The final exam will be held on-campus. It will be closed book and will count for 80% of your final grade. You will be allowed to have one A4 sheet (both sides) of handwritten notes.

  • Announcements Forum
  • Student Forum External tool
• 

  Indicative Class Schedule

  21.02.2023	Lecture 1: Introduction, Digital Images, K-Nearest Neighbors Exercise Session: Introduction to Python
  28.02.2023	Lecture 2: K-Nearest Neighbors, K-Means Clustering Exercise Session: Introduction to NumPy
  07.03.2023	Lecture 3: Clustering, Linear Regression Exercise Session: K-Nearest Neighbors
  14.03.2023	Lecture 4: Linear Classification, Max-Margin Classifiers Exercise Session: K-Means Clustering
  21.03.2023	Lecture 5: Optimization Basics Exercise Session: Logistic Regression
  28.03.2023	Lecture 6: Adaboost and Support Vector Machines Exercise Session: Project MS1
  04.04.2023	Lecture 7: Support Vector Machines Exercise Session: SVM
  11.04.2023	Easter Break
  18.04.2023	Lecture 8: Multi Layer Perceptron Exercise Session: Project MS1
  25.04.2023	Lecture 9: Multi Layer Perceptron (Part 2) Exercise Session: MLP
  02.05.2023	Lecture 10: CNN (Part 1) Exercise Session: CNN (part 1)
  09.05.2023	Lecture 11: CNN (Part 2) Exercise Session:  UNet
  16.05.2023	Lecture 12: Linear Dimensionality Reduction Exercise Session: Linear dimensionality reduction (PCA, LDA)
  23.05.2023	Lecture 13: Non-linear Dimensionality Reduction Exercise Session: Non-linear Dimensionality Reduction + Project MS2
  30.05.2023	Lecture 14: Review Session Exercise Session: Project MS2
• 

  Reference Text Books

  C.M. Bishop, Pattern Recognition and Machine Learning, Springer, 2006.

  M. Welling, A First Encounter with Machine Learning, 2011

  
  
• 

  Introduction to the Class

  • Introduction File
  • Recorded Lecture 20/21: Introduction and kNN URL
  • Python and NumPy Primer File
  • Python Cheatsheet File
  • Matplotlib Cheatsheet File
  • Recorded Lecture 20/21: Python and NumPy Primer URL
  • Exercise 1: Introduction to Python File
  • Exercise 1: Introduction to Python Solutions File
  • Exercise 2: Introduction to NumPy File
  • Exercise 2: Introduction to NumPy Solutions File
• 

  K-Nearest Neigbors and K-Means

  • K-Nearest Neighbors File
  • Recorded Lecture 20/21: kNN URL
  • Exercise 3: kNN File
  • Exercise 3: kNN Solutions File
  • K-Means Clustering File
  • Recorded Lecture 20/21: K-Means URL
  • Exercise 4: K-Means File
  • Exercise 4: K-Means Solutions File
• 

  Linear Regression and Classification

  • Linear Regression File
  • Linear Classification File
  • Recorded Lecture 20/21: Linear Classification (1) URL
  • Recorded Lecture 20/21: Linear Classification (2) URL
  • Exercise 5: Logistic Regression File
  • Exercise 5: Logistic Regression Solutions File
  • Max Margin Classifiers File
  • Recorded Lecture 20/21: Max-Margin Classifiers URL
• This week
  

  Non Linear Classification

  • AdaBoost File
  • Support Vector Machines File
  • Proof of Cover's Theorem (optional) File
  • Recorded Lecture 20/21: Adaboost URL
  • Recorded Lecture 20/21: Linear SVMs URL
  • Recorded Lecture 20/21: Kernel SVMs URL
• 

  Non Linear Optimization

  • Optimization Basics File
• 

  Project

  • Project Description Document URL

    Last edit: 18.03.2023

  • Dataset MS1: HASYv2 URL
  • Framework MS1 File
  • Milestone 1 Submission Link - Deadline: April 21st 23:59 Assignment
• 

  Mock Exam

  • Mock Exam File
  • Mock Exam Solutions File