Psych 625: Applied Machine Learning
Morteza Dehghani
Fall 2022
Office Hours: Tues 10-12 Class Hours: Th 2-6pm
Office: SGM 607 Class Room: VHE210
Course Description
In the past decade, social scientists have been facing a quantitative change in technology. This
change can be summarized in two main points: 1. availability of vast and seemingly insurmount-
able volumes of human-related data, and 2. constantly increasing computational power. These
have provided an unprecedented opportunity to study and model human cognition with range
and detail previously not imaginable. Moreover, there is growing interest (e.g. in marketing) to
use such data for predicting a variety of human behavior. Applied Machine Learning focuses on
methods in computer science, specifically in machine learning, which can help us achieve these
outcomes. This course is followed by Psych 626: Text as Data which focuses on the applications
natural language processing, guided by psychological theories, for identifying various social and
cognitive properties evident in human related big data.
The intended audience for this course is psychology graduate students, and more broadly
graduate students in social sciences, who are interested in using machine learning techniques
for analysis of data, Also, this course may be of interest to PhD students in communications,
computer science and the business school.
Learning Objectives
This course is designed to be hands-on and students are expected to learn how to apply different
machine learning techniques for analyzing different types of data. In order to achieve this objec-
tive, each discussed topic is accompanied by a lab session in which we examine how to use that
technique on a data set. Homework review sessions are used for helping students troubleshoot
their code and also for going over the homework.
• Prerequisite(s): Instructor permission.
• Recommended Preparation: Psych 501 or a similar introductory statistics course.
1
Applied ML
Course Notes
Lecture notes and homework assignments will be posted on Blackboard. Students are also highly
encouraged to use the course forum on Blackboard.
Technological Proficiency and Hardware/Software Required
This class includes lab sessions. Students are required to bring a laptop to class. Homework
assignments are programming problems that need to be written in R.
Required Books
• Gareth, J., Daniela, W., Trevor, H., and Robert, T. (2021). An introduction to statistical learning:
with applications in R (Second Edition). Spinger
Description and Assessment of Assignments
1. Homework assignments. Each week students will complete programming problems from
one of the required books. The assignments will be graded based on both output and style
of the code. The homework material will be reviewed during homework review sessions.
2. Lab presentation. Each student will do a lab presentation in which a particular lab module
is taught to others using a different dataset than the one used in the book. This dataset will
be made available to the class by the second week of class. By having enrolled in the class
all student acknowledge the copyright information regarding this dataset.
3. Class Projects. Students will complete five class projects. These projects will be relatively
heavy programming assignments requiring students to use R to implement some specific
statistical technique. The first project will be relatively easy and not time consuming. The
other projects, however, will take substantial time.
Grading Policy
• 25% Homework
• 5% Project 1
• 20% Project 2
• 15% Project 3
• 10% Project 4
• 15% Project 5
• 5% Presentations
• 5% Participation
2/6
Applied ML
Assignment Submission Policy
Homework will be assigned on Thursdays and will be due the following Thursday at 11am,
before the start of class submitted on Blackboard. Usually, three questions will be assigned for
homework, and you have the option of answering two of them. All homework turned in any
later than 11:10am will be considered late. Students will be allowed a total of seven late days that
can be used on the assignments. In exceptional circumstances, arrangements must be made in
advance of the due date to obtain an extension. Once you have used up your seven late days, one
additional day late will result in a 25% reduction in the total score, two additional days late will
yield a 50% reduction, and no credit will be given for three or more additional days late. Late
days are in units of days, not hours, so using up part of a day uses up the whole day. The final
project report, plus the R code used, will be due on the day of the final exam. All assignments,
including the projects, need to be written using knitR. Copied and pasted code/results will not
be accepted.
Schedule and weekly learning goals
The schedule is tentative and subject to change.
Week 01, 08/25: : Introduction, Statistical Learning & Linear Regression
• ISLR Chapter 2
• ISLR Chapter 3
• Lab: Linear Regression (ISLR 3.6)
• Intro to knitR package
• HW 1; Project 1 assigned
Week 02, 09/01: Classification
• ISLR Chapter 4
• Bayesian Classifiers
• Lab: Logistic Regression, LDA, QDA and KNN (ISLR 4.6)
• HW 1 due; HW 2 assigned
• Review of HW 1
Week 03, 09/08: Resampling Methods
• ISLR Chapter 5
• Lab: Resampling Methods (ISLR 5.3)
• HW 2 due; HW 3 assigned
3/6
Applied ML
• Project 1 due; Project 2 assigned
• Review of HW 2
Week 04, 09/15: Linear Model Selection
• ISLR Chapter 6
• Lab: Regularization (ISLR 6.5, 6.6 & 6.7)
• HW 3 due; HW 4 assigned
• Review of HW 3
Week 05, 09/22: Moving Beyond Linearity
• ISLR Chapter 7
• Lab: Moving Beyond Linearity (ISLR 7.8)
• HW 4 due; HW 5 assigned
• Review of HW 4
Week 06, 09/29: Tree-based Methods
• ISLR Chapter 8
• Lab: Decision Trees (ISLR 8.3)
• Project 2 due; Project 3 assigned
• HW 5 due; HW 6 assigned
• Review of HW 5
Week 07, 10/06: Support Vector Machines
• ISLR Chapter 9
• Support Vector Regression (Handouts)
• Lab: Support Vector Machines (ISLR 9.6)
• HW 6 due; HW 7 assigned
• Review of HW 6
Week 08, 10/13: Fall Recess
4/6
Applied ML
Week 09, 10/20: Neural Networks & Deep Learning I
• The Perceptron (Handouts)
• ISLR Chapter 10
• HW 7 due;
• Review of HW 7
Week 10, 10/27: Neural Networks & Deep Learning II
• Introduction to keras
Week 11, 11/03: Neural Networks & Deep Learning III
• Lab: Deep Learning (ISLR 10.9)
• HW 8 assigned
• Project 3 due; Project 4 assigned
Week 12, 11/10: Unsupervised Learning
• ISLR Chapter 12
• HW 8 due; HW 9 assigned
• Review of HW 8
• Lab: Unsupervised Learning
Week 13, 11/17: Thanksgiving Holiday
Week 14, 11/24: Probability calculus, Bayesian networks: syntax and semantics
• From Propositional to Graded Beliefs (Handouts)
• Updating Beliefs (Handouts)
• Capturing Independence Graphically (Handouts)
• Parameterizing the Independence Structure (Handouts)
• HW 9 due; HW 10 assigned
Week 15, 12/01: Bias in ML
• Barocas, Solon, Moritz Hardt, and Arvind Narayanan. "Fairness in machine learning." Nips
tutorial 1 (2017): 2. Available at: https://fairmlbook.org/. Chapters 1, 2 &4
5/6
Applied ML
Statement on Academic Conduct and Support Systems
Academic Conduct
Plagiarism — presenting someone else’s ideas as your own, either verbatim or recast in your own
words — is a serious academic offense with serious consequences. Please familiarize yourself
with the discussion of plagiarism in SCampus in Section 11, Behavior Violating University Standards
https://scampus.usc.edu/1100-behavior-violating-university-standards-and-appropriate-
sanctions/. Other forms of academic dishonesty are equally unacceptable. See additional in-
formation in SCampus and university policies on scientific misconduct, http://policy.usc.edu/
scientific-misconduct/.
Discrimination, sexual assault, and harassment are not tolerated by the university. You are
encouraged to report any incidents to the Office of Equity and Diversity http://equity.usc.edu/
or to the Department of Public Safety http://capsnet.usc.edu/department/department-public-
safety/online-forms/contact-us. This is important for the safety whole USC community. An-
other member of the university community — such as a friend, classmate, advisor, or faculty
member — can help initiate the report, or can initiate the report on behalf of another person. The
Center for Women and Men http://www.usc.edu/student-affairs/cwm/ provides 24/7 confiden-
tial support, and the sexual assault resource center webpage
[email protected] describes reporting
options and other resources.
Support Systems
A number of USC’s schools provide support for students who need help with scholarly writing.
Check with your advisor or program staff to find out more. Students whose primary language
is not English should check with the American Language Institute http://dornsife.usc.edu/ali,
which sponsors courses and workshops specifically for international graduate students. The Of-
fice of Disability Services and Programs http://sait.usc.edu/academicsupport/centerprograms/
dsp/home_index.html provides certification for students with disabilities and helps arrange the
relevant accommodations. If an officially declared emergency makes travel to campus infeasible,
USC Emergency Information http://emergency.usc.edu/will provide safety and other updates,
including ways in which instruction will be continued by means of blackboard, teleconferencing,
and other technology.
IMPORTANT: COVID-19 PROTOCOLS
Students must comply with all COVID-19 safety protocols outlined by federal, state, local, and
university policies. These policies will likely evolve with the changing conditions of the COVID-
19 pandemic and may include social distancing, the use of face coverings at all times, proof of
vaccination, and regular COVID testing, among others. Depending on the policies outline by the
above authorities, and the conditions of the class, the class might switch between meeting online
and in person.
6/6
