COMP 551 Applied Machine Learning

Instructor: Sarath Chandar

Applied Machine Learning Sarath Chandar

Instructor

Sarath Chandar

2

Applied Machine Learning Sarath Chandar

Lead Teaching Assistant

3

Prasanna Parthasarathi

Applied Machine Learning Sarath Chandar

Teaching Assistants

4

Ali Emami Scott Fujimoto Carlos Gonzalez Gandharv Patil

Edward Smith Xin Tong Wang Nadeem Ward

Applied Machine Learning Sarath Chandar

Course Details

Website: http://sarathchandar.in/teaching/2018/fall/comp551/

Schedule: http://sarathchandar.in/teaching/2018/fall/comp551/schedule.html

Schedule will be updated regularly! So it is always tentative!

5

Applied Machine Learning Sarath Chandar

What is this course about?

This is an introductory course in Machine Learning which covers fundamental topics in supervised learning and unsupervised learning.

6

Applied Machine Learning Sarath Chandar

Course Structure

● 2 lectures per week.● Weekly reading.● In-class surprise quizzes.● 1 in-class end-term exam.

7

● 3 tutorial sessions in total.● 3 programming assignments.● 1 Kaggle competition.● 1 final project.

Theory Practice

Applied Machine Learning Sarath Chandar

Grading Info

● In-class surprise quizzes - 10%● 3 programming assignments (individual) - 25%● Kaggle competition (team of 3*) - 15%● Course Project (team of 3*) - 20%● Written Exam - 30%

● You cannot have same team members for both competition and project.

8

Applied Machine Learning Sarath Chandar

Programming Assignments

● 3 programming assignments.● Individual assignments.● Submission instructions will be released with the assignments.● Late work will be subject to a 30% penalty, and can be submitted up to 3 days after the

deadline.● We will use gradescope for all the course related submissions.● You have to use Python 3 for all the three assignments.● Tutorial-1 will cover introduction to python 3.

○ Date: September 14○ Time: 6 pm to 7:30 pm○ Venue: ENGMC 304

9

Applied Machine Learning Sarath Chandar

Code of Conduct

● Zero tolerance for plagiarism and cheating!● Each assignment/project/competition comes with rules and you MUST follow them.

10

Applied Machine Learning Sarath Chandar

Prerequisites

● Knowledge of a programming language.● Knowledge of probabilities/statistics (e.g. MATH-323, ECSE-305).● Knowledge of calculus and linear algebra.● Some AI background is recommended (e.g. COMP-424, ECSE-526) but not required.● First week’s reading covers some prerequisite materials.

● Assignment-0 covers some problems based on the prerequisites. This is NOT graded. But you must submit them to get used to Gradescope.

● Assignment-0 is already out!

11

Applied Machine Learning Sarath Chandar

Discussions

● We will use reddit for class discussions.● Check myCourses for link to the subreddit for the class.● You can post your questions lecture-wise.● Please do not create a new link or a post. Use the posts assigned for corresponding

lecture.● Sign up before next class!● If you are not in myCourses, email Prasanna with your first name, last name,

McGill ID, and your reddit user name.

12

Applied Machine Learning Sarath Chandar

Course Feedback

● You can submit your feedback about the course at any point of time.● Check myCourses for a link to the Google form. (Also available in Reddit)● Feel free to give both positive and negative feedback about every lecture!

13

Applied Machine Learning Sarath Chandar

Surprise Quizzes

● Any number of surprize quizzes.● Always bring an A4 sheet and pen to the lectures.

14

Applied Machine Learning Sarath Chandar

Course calendar

● Course calendar available in the website.● Check the calendar for

○ Office hours location and time.○ Tutorials location and time.○ Submission deadlines.

15

Applied Machine Learning Sarath Chandar

Email Policy

● All your emails about sick leave, extensions, complaints should be directed to Prasanna only. Email me only if it is very personal.

● Do NOT email the TAs for course doubts or assignment doubts. You doubts should be posted only in Reddit forum.

● Do NOT email the TAs for regrading your assignments. Assignment regrading will be done only through Gradescope.

16

Applied Machine Learning Sarath Chandar

Tips to excel in the course..

● Read the lecture notes of the previous class before coming to the class.● Finish all the readings that are assigned every week.● Do all the self-assessment questions posted for every lecture.● Attempt to solve all the non-graded assignments too.

17