Machine learning programming

UPDATE! Please attend the course in person in the room corresponding to your last name's first letter.

CM 1 111/CM 1 112 : last names starting with letters A to D

GR B0 01: last names starting with letters E - L

GR C0 02: last names starting with letters M - Z

==========================================================================

Time and Location: The course is given each Wednesday 8h15-10h00.

INCLUSIVENESS: to be inclusive of all students, and in particular those who cannot be on campus for health reason, the Machine Learning Programming class will be given both on-site in room  GR B0 01, GR C0 02, CM 1 111 and on-line through discord simultaneously (zoom will be used for the first 15 minutes to introduce the topic, or a video will be provided on Moodle).

Content and Programming language: This is a programming class with a total of 6 graded programming assignments. Programming is done in MATLAB. The class will request you to program the following algorithms as assignments:

• Principal Component Analysis (PCA)
• K-means
• K-nearest neighbour (KNN)
• Gaussian Mixture Model (GMM)
• Applications of GMMs (Clustering, Classification, Regression)
  
• Neural Networks

Each of those assignments consists of two course sessions and have to be submitted the day before the next assignment starts.

=======================================================

Prerequisites: The material for the assignments follows the material given in the applied machine learning course. It is not mandatory to follow the applied machine learning course. However, students who take the machine learning programming course must know the machine learning algorithms listed above.

Videos of theoretical material: Videos of the materials taught in Applied Machine Learning course and necessary for this course are available on the moodle page of the Applied Machine Learning course, as well as through a dedicated SwitchTube Channel.

Lecture Notes

The Lecture Notes for the Applied Machine Learning course can be downloaded by following this link. These can serve as support for the theoretical content behind the algorithms we will implement in the MLP course. For information on the programming exercises, you must refer to the slides and assignment sheet of each week.

=======================================================

Grading: 

Grading is based on assignments to be submitted regularly during the semester and a mid-term exam, given on December 4 during the first Class hour: 8h15-9am.

Assignments are graded on an individual basis. In other words, each student submit his/her own assignment. Assignments will be checked for plagiarism. Assignments must be turned it in by the deadline (check deadline on each assignment). 1pt (out of the full grade) will be removed for each day late. A day late starts one hour after the deadline. There are 6 graded assignments during the whole course. The final assignment grade is the average grade of all the 5 best grades out of 6 graded assignments (leave one out).

Grade from average of assignments is worth 75% of the total grade. Midterm exam on December 4 is worth 25% of the total grade.

======================================================================

Teaching Assistants

Harshit Khurana: harshit.khurana@epfl.ch and Aubane Lachat In GR C002

Luyin Hu luyin.hu@epfl.ch and Paolo Giaretta  In GR B001

Sthithpragya Gupta and Nina Lahellec In CM 1 111/CM 1 112

Responsible Instructor

email: aude.billard@epfl.ch

• Introduction to Machine Learning Programming course
• MATLAB and Machine Learning proficiency exercise session 1
  
• Introduction given by Kunpeng Yao
  

• MATLAB and Machine Learning proficiency exercise session 2

• Review of Principal Component Analysis 
• Implementation of Principal Component Analysis (PCA)
• Introduction given by Harshit Khurana
• Optional: view video on derivation of PCA to refresh your memory, if needed.
  

• Implementation of PCA
• Application and analysis of PCA

• Review of k-Means clustering algorithm
• Implementation of k-Means
• Introduction given by Kunpeng Yao
• Optional: View video on K-means to refresh your memory if needed
  

• Review on Metrics for k-Means evaluation
• Analysis and Evaluation of k-Means

• Review of k-NN classification algorithm
• Implementation of k-NN
• Introduction given by Yang Liu
• Optional: View video on kNN to refresh your memory if needed

• Review of Gaussian Mixture Modeling (GMM)
• Implementation of GMM

• Applications of GMM for classification, regression and sampling
• Introduction given by 

• 8h15-9am: Exam in programming room - multiple choice - worth 25% total course grade taking place in rooms CO020-CO021-CO023, BC07-08 (see students' room allocation)
  
• 9h15-10am: Programming Session Continued - Applications of GMM for classification, regression and sampling. Students must go back to the rooms used for the class.
  

• Classification with Neural Network
  
• Introduction given by David Gonon