CMPUT 466/566 (Fall 2024) Syllabus
Machine Learning Essentials

Suggestions for contacting the instructor and TAs

Lecture time and classroom:

T, Th 12:30PM - 1:50PM, Sep 3 - Dec 9

CCIS L1-160

● No course activities during the reading week

Online lecture hall: https://meet.google.com/vqo-vkxv-osy

Dial-in: (US) +1 337-573-0059 PIN: 584 572 100#

• Office hours start from the second week. No office hours on statutory holidays and during the reading week.
• If a student wishes to make an appointment with a TA (10min each slot), they will send an email to the TA before the date.
• Office hour schedule may be changed depending on the need.

Students are encouraged to reach out to the instructor if TAs’ answer is not satisfactory.

The instructor and TAs will not answer assignment-related questions before the solution is released.

COURSE CONTENT

Course Description:

Machine learning teaches a machine to learn from previous experience and makes a prediction for (possibly new) data. This course covers standard materials of a “Machine Learning” course, such as linear regression, linear classification, as well as non-linear models. In the process, we will have a systematic discussion on issues such as training criteria, inference criteria, bias-variance tradeoff, etc. The goal of the course is to build a solid foundation of machine learning; so there would be intensive math derivations in lectures, assignments, and exams.

Please fulfill the departmental requirements.

The department asks instructors normally not to waive prerequisites.

By the end of this course, the student will understand the foundations of machine learning and gain experience in machine learning applications.

Official textbook: Bishop, Pattern Recognition and Machine Learning.

The instructor will provide lecture notes, which may also suffice. If not, please use the above text book. [survey on textbooks]

References: link

Tentative topic list:

Linear regression

• Mean squared error (as heuristics)
• Closed-form solution
• Gradient descent
• Maximum likelihood estimation
• Maximum a posteriori training
• Bias-variance tradeoff
• Train-validation-test framework

Linear classification

• Discriminative model: Logistic regression
• Multi-class softmax
• Maximum a posteriori inference
• Generative model: Naïve Bayes
• Discriminant model: Linear SVM (bonus lecture)

Nonlinear models

• Neural networks
• Kernels methods: Non-linear SVMs (bonus lecture)

Note: The actual lecture pace may vary depending on students’ background and interest. Exams will be based on main lectures only. Bonus lectures are not required.

LEARNING RESOURCES

The Academic Success Centre provides professional academic support to help students strengthen their academic skills and achieve their academic goals. Individual advising, appointments, and group workshops are available year round in the areas of Accessibility, Communication, Learning, and Writing Resources. Modest fees apply for some services.

Faculty of Science Student Services:

GRADE EVALUATION

Students must verify this date on BearTracks when the Final Exam Schedule is posted.

Grades are unofficial until approved by the Department and/or Faculty offering the course.

No collaboration is allowed for a basic mini-project (or any assignment). For a non-trivial project, collaboration may be allowed up to 3 students. In this case, the students have to form a group themselves and apply in the notice of intent (NOI). Each of the team members MUST have substantial and similar previous machine learning background. The team approval will be based on students’ previous experience and the intended topic.

The mid-term exam is optional. When the letter grade is computed, mid-term marks =max{mid-term percentage, final percentage} * 30

Exams are in-person. In exceptional circumstances, the instructor may approve remote exams (video/audio on, screen-sharing).

• Reasonable accommodations include: being sick, being in quarantine, attending academic conferences, remote attendance as a non-UofA student
• Pre-disapproved excuses: feel lazy, want to cheat at home, personal vacation plan, etc.
• If a student sees a need for remote exams, please discuss with the instructor as soon as possible. Please do not arrange your travel plan until you are approved by the instructor.

[Example exam conduct]

• Bonus for earnest efforts towards course objectives and expected learning outcomes: Students should make earnest efforts towards the assessments that are linked with course objectives and expected learning outcomes.

When students make unscrupulous efforts to gain marks, it will hurt the course objectives and expected learning outcomes.

A student may get up to 5 bonus marks based on the instructor’s discretion. Not earnest efforts include

- Submitting late assignments

- Arguing marks and grades without a scientific reason

- with a scientific reason is okay

- Violating exam policies (e.g., solving problems before the exam starts or failure to stop writing by the end of the exam)

The bonus marks are calculated based on the following schedule: 2, 1, 1, 0.5, 0.25, 0.125, etc.

If the instructor grants bonus marks, the student should send a request to the course email to claim the bonus points. The bonus mark request must be sent on the same day when the bonus mark is earned, because the instructor may not remember the details of past lectures.

Note: AI-flavored writing demonstrates poor presentation skills. For example, the text is oftentimes grandiose but empty. This will result in a devastatingly low mark (including a mark of 0) by merit, regardless of whether there is proof of using AI tools.

Statement of Confidentiality: All course materials, and their direct derivatives (e.g., assignment solutions), are confidential. Students should not publish or share the lectures, notes, assignments, solutions, and exams of the course.

Misrepresentation of Facts to gain a re-examination is a serious breach of the Student Academic Integrity Policy.

Letter grade:

The final letter grade will be given by some cut-off based on numerical marks. Assuming a student does reasonably well in all assignments and the course project, then the letter grades roughly maps to the following criteria:

A = The student well understands lecture materials and can generalize to certain new problems

A- = The student well understands lecture materials but is unable to generalize to new problems

B+ = The student understands most part of the lecture materials, but a few details are missing

B = The student understands some part of the lecture materials, but several parts are missing

B- = The student understands some part of the lecture materials, but a significant portion is missing

Letter grade cutoffs may be different for undergrads and undergrads, because they are two courses.

In the past, the instructor typically sent out 30-40% A or A+. Please disregard eClass course total, as it is not correctly set up. We will calculate course total marks on a spreadsheet strictly following the syllabus.

POLICIES FOR LATE AND MISSED WORK

Written assignments. Every submission is due in Edmonton time, but is granted a two-day free extension. The student should have finished and submitted the solution by the (first) deadline. The free extension is intended to resolve all issues, including the confusion of time zones, temporary computer/internet/power failure, and any other personal emergencies from the student. Further extensions will not be granted. Inquiring or submitting assignments after the second deadline is considered unscrupulous efforts towards learning objectives and expected course outcomes. 

In addition to free extensions, we automatically approve an excused absence (EA) if a student applies before the (first) deadline to the course email. No explanation is needed.

Coding assignments. Coding assignments will be given enough time and the deadline will be extended as well. However, coding assignments cannot be EA’ed, because the exam does not test coding aspects. Failure to submit the coding assignment solution to eClass means a mark of zero for the coding assignment.

Inquiring or submitting assignments after the second deadline reflects unscrupulous efforts towards course objectives and expected learning outcomes.

Mid-term exam is optional, because the mid-term will be lifted up to the mark percentage of the final.

Please note that EA’ing everything to the final exam is unwise, because FoS will only approve deferred final exams (see below) when a significant amount of coursework is done. A denial of the deferred final means the EA’ed marks are 0.

Missed Term Work or Final Exam Due to Non-medical Protected Grounds (e.g., religious beliefs):

It is imperative that students review the dates of all course assessments upon receipt of the course syllabus, and apply AS SOON AS POSSIBLE to ensure the timely application of the accommodation. Students who apply later in the term may experience unavoidable delays in the processing of the application, which can affect the accommodation.

Deferred Final Examination:

A student who cannot write the final examination due to incapacitating illness, severe domestic affliction or other compelling reasons can apply for a deferred final examination. Such an application must be made to the student’s Faculty office within two working days of the missed examination and must be supported by appropriate documentation or a Statutory Declaration (see calendar on Attendance). Deferred examinations are a privilege and not a right; there is no guarantee that a deferred examination will be granted. The Faculty may deny deferral requests in cases where less than 50% of term work has been completed. Misrepresentation of facts to gain a deferred examination is a serious breach of the Student Academic Integrity Policy.

STUDENT RESPONSIBILITIES

Students are particularly urged to familiarize themselves with the provisions of the Student Academic Integrity Policy and the Student Conduct Policy, and avoid any behaviour that could potentially result in suspicions of academic misconduct (e.g., cheating, plagiarism, misrepresentation of facts, participation in an offence) and non-academic misconduct (e.g., discrimination, harassment, physical assault). Academic and non-academic misconduct are taken very seriously and can result in suspension or expulsion from the University.

The College of Natural and Applied Sciences (CNAS) has created an Academic Integrity for CNAS Students eClass site. Students can self enroll and review the various resources provided, including the importance of academic integrity, examples of academic misconduct & possible sanctions, and the academic misconduct & appeal process. They can also complete assessments to test their knowledge and earn a completion certificate.

"Integrity is doing the right thing, even when no one is watching." -- C.S. Lewis

Appropriate Collaboration:

Collaboration for projects is subject to the instructor’s approval, and such applications must be addressed in notice of intent (NOI) before the NOI deadline.

Exam Conduct:

Students Eligible for Accessibility-Related Accommodations:

It is recommended that students apply AS SOON AS POSSIBLE in order to ensure sufficient time to complete accommodation registration and coordination. Students are advised to review and adhere to published deadlines for accommodation approval and for specific accommodation requests (e.g., exam registration submission deadlines). Students who request accommodations less than a month in advance of the academic term for which they require accommodations may experience unavoidable delays or consequences in their academic programs, and may need to consider alternative academic schedules.

Learning and Working Environment:

If you are experiencing harassment, discrimination, fraud, theft or any other issue and would like to get confidential advice, please contact any of these campus services:

It's normal for us to have different mental health experiences throughout the year. Know that there are people who want to help. You can reach out to your friends and access a variety of supports available on and off campus at the Need Help Now webpage or by calling the 24-hour

Student Self-Care Guide:

This Self-Care Guide, originally designed by the Faculty of Native Studies, has broader application for use during students’ learning. It provides some ideas and strategies to consider that can help navigate emotionally challenging or triggering material.

Land Acknowledgement:

To learn more about the significance of this land acknowledgement, please read this useful article and associated links to more information.