MAT 3373 (Fall 2020)

Methods of Machine Learning

Professor: A. Smith
Office : 150 Louis Pasteur (STEM), Room 537

Official Course Times: Tuesday from 1-2:30 PM and Thursday from 11:30 AM-1 PM.

Urgent Suggestion: If you feel that you need accommodations, you should
contact Access Services as soon as possible. Accommodations are not limited
to those with physical disabilities. Unfortunately there is a deadline to receive
accommodations, so please do not delay.

Grading: Your grade in the course will be calculated as follows:

Assignments 65%
Midterm 10%
Final Exam 25%

ˆ The assignments are of very different length. I will take your average grade
per question rather than per assignment. To give a concrete example: if
there were two assignments, and on HW1 you got 1 of 3 questions right
while on HW2 you got 9 of 10 questions right, your HW grade would be
9+1 1 1 9
10+3 ≈ 0.77, not 2 ( 3 + 10 ) ≈ 0.62.

ˆ Assignments will not be accepted after solutions have been posted. I

expect to post solutions quite soon after assignment deadlines, though
delays of a few days are possible (if e.g. I find an error in my solution,
think that a student solution is particularly clever and worth sharing, or
am unable to connect for whatever reason). I will typically accept “late”
homework sent to me before I post solutions, but of course sending things
in more than a few minutes late is a bit of a gamble. Note: Just because
one late homework is accepted does not mean that further late homework
will be accepted!

1
 ˆ There will be no extra make-up midterm. If you miss the midterm for an
allowed reason, the midterm grade will be replaced with the final exam
grade.
ˆ Due to various personal emergencies, you may not be able to complete all
homework assignments. To partially account for this, on each assignment
I will drop the lowest question grade with no questions asked. If you need
further accommodations, you will need to provide some justification and
likely some documentation. See first day of class for further details.
ˆ If you are unable to write the final exam, then the policies of the Faculty
will apply.

Textbook: We will be using a variety of resources, all of which should be posted

on Brightspace or otherwise sent to you. You don’t need to buy any textbooks.

Office Hours and Other Discussion:

I encourage students to ask questions (or open discussions) throughout the
term. Please don’t feel that you can only ask questions about specific homework
problems - if you feel that e.g. you are missing some foundations, we can work
together on that as well.
Here are some venues for asking questions:

1. Questions about math, statistics and computation are welcome (and very
much encouraged) during class times.
2. For all text communication besides personal accommodations and office
hour requests I have set up a Piazza forum at:

piazza.com/uottawa.ca/winter2021/mat3373

You should ask all short questions about course content, logistics, and so
on in that forum, and I will respond to personal emails by asking you to
repost there. I will check Piazza 5 or more times per week throughout
the term. Note that Piazza has some other helpful features (e.g. it is also
possible to set up polls there, and you can write posts without using your
name).
3. If you need any accommodations, please start by writing to the Student
Accommodations office if appropriate. If not, or if they are not helpful,
please email me directly and include enough information for me to under-
stand why they are not the right choice. In either case, please do this as
soon as possible - it is vastly easier to find an appropriate accommodation
with more time to work with. I don’t want to make this course un-
necessarily stressful, but can’t always guess as to what is causing
problems. Please tell me!

2
 4. Since meeting in person is discouraged, I will replace traditional office
hours by Zoom videoconferencing meetings. To set one up, please send
me an email with (i) a list of times that you are available and (ii) a quick
summary of what you’d like to discuss. I’ll try to accomodate any requests,
but the chance of finding a common meeting time goes up if you send a
request earlier and list more meeting times. In particular it is hard to
schedule in ¡24 hours and/or with a single possible meeting time. Details
may change during term if this approach is not going well; I will update
via Brightspace.

Official Description of the Course : Multivariate linear and polyno-

mial regression, logistic regression, k-nearest neighbours. Regularization meth-
ods, Bayesian inference, naive Bayes classification. Cross-validation, bootstrap.
Tree-based methods. Principal component analysis. Clustering.

Prerequisite: MAT 1341, and either MAT2377 or both of MAT 2371, MAT
2375.

Homework: Homework will be available on Brightspace, and deadlines are

also listed on Brightspace when they are posted. Homework can be completed
and submitted early (and I strongly suggest handing in homework a little early
to avoid surprises).
Other accommodations are available for those who need them, if you write
to me sufficiently far in advance that they can be implemented.

Midterm: The midterm will be held in class time on February 25. You may
use your notes and a Python or R programming environment.

FINAL EXAM : The exam of three hours will be held during the examination
period and will cover all subject matter seen during the semester. It is your
responsibility to determine the date, place and time of the final exam. Do not
rely on your friends. An error from them could cause you many troubles. In
addition, never write an examination, especially a final exam, if you are sick. We
can not take this into account after the fact. Check the regulations on this issue
in the regulations of the Faculty of Sciences, see https://science.uottawa.
ca/en/students-etudiants/examens.html.

Academic Fraud:
Academic Regulation 14 defines academic fraud as “any act by a student

3
that may result in a distorted academic evaluation for that student or another
student. Academic fraud includes but is not limited to activities such as:
1. Plagiarism or cheating in any way;
2. Submitting work not partially or fully the student’s own, excluding prop-
erly cited quotations and references. Such work includes assignments,
essays, tests, exams, research reports and theses, regardless of whether
the work is written, oral or another form;
3. Presenting research data that are forged, falsified or fabricated;
4. Attributing a statement of fact or reference to a fabricated source;

5. Submitting the same work or a large part of the same piece of work in more
than one course, or a thesis or any other piece of work submitted elsewhere
without the prior approval of the appropriate professors or academic units;
6. Falsifying or misrepresenting an academic evaluation, using a forged or
altered supporting document or facilitating the use of such a document;
7. Taking any action aimed at falsifying an academic evaluation.”

Notice against sexual violence

The University of Ottawa will not tolerate any act of sexual violence. This
includes acts such as rape and sexual harassment, as well as misconduct that
take place without consent, which includes cyberbullying. The University, as
well as various employee and student groups, offers a variety of services and
resources to ensure that all uOttawa community members have access to con-
fidential support and information, and to procedures for reporting an incident
or filing a complaint.

Drop Date: March 26. See uOttawa academic calendar for other important
dates.

Note that this syllabus is subject to modification. Modifications

will be announced in class and/or on Brightspace.