Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Spring 2024

CMPE 597 Sp. Tp. Deep Learning

Instructor: İnci M. Baytaş

Email: inci.baytas@boun.edu.tr
Website: http://cmpe.boun.edu.tr/~incibaytas
Office: BM 24
Office Hours: Appointment by email

Time: FFF 567

Location: BM A3

Textbooks:

– Simon J.D. Prince, Understanding Deep Learning, MIT Press Dec 5th 2023.
https://udlbook.github.io/udlbook/
– I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning, MIT Press, 2016.
https://www.deeplearningbook.org/

Prerequisites
Basic knowledge of machine learning, linear algebra, probability, statistics, and optimization is
required to succeed in this course. If you are not familiar with machine learning, taking/auditing
a machine learning class is encouraged before taking this course. Python will be used in the
course projects. Students will need to use Tensorflow or Pytorch in the course projects.

Course Description
An introduction to artificial neural networks, deep learning, fundamental deep architectures,
and their applications.

Objectives and Learning Outcomes


The objective of this course is to provide the foundations of neural networks and introduce
common deep learning models. Upon successful completion of the course, students are expected
to gain an understanding of the following:

– fundamentals of learning from data paradigm


– perceptron and multilayer perceptron
– back-propagation algorithm
– optimization techniques in deep learning

1
– training neural networks, parameter tuning
– deep network architecture design
– convolutional neural networks, recurrent neural networks, auto-encoders, and deep gener-
ative models
– some popular applications of deep models.

Course Topics

– Multilayer Perceptron
– Training Deep Networks
– Convolutional Neural Networks
– Sequence Modeling, Recurrent Neural Networks, Transformer Networks,
– Graph Neural Networks
– Auto-Encoders
– Deep Generative Models, Variational Auto-Encoders, Generative Adversarial Networks,
Diffusion Models

Assessment

– Quiz (3) 15%


– Midterm 30%
– Project 20%
– Final 35%

Tentative Class Schedule:

Week Date Topic Event


1 February 16 Introduction
2 February 23 Multilayer Perceptron Project proposals due
3 March 1 Multilayer Perceptron
4 March 8 Training Deep Networks Quiz 1
5 March 15 Training Deep Networks
6 March 22 Convolution Neural Networks Quiz 2
7 March 29 Convolution Neural Networks
8 April 5 Sequence Modeling Midterm
9 April 12 Spring Break
10 April 19 Sequence Modeling + Graph Neural Networks Project progress discussion
11 April 26 Graph Neural Networks + Auto-Encoders
12 May 3 Deep Generative Models Quiz 3
13 May 10 Deep Generative Models
14 May 17 Project Presentations Project reports due

2
Class Participation: Please do not take this class if you have conflicts in your schedule.
Attendance will be taken on exam days. Please consider this before you register.

Course Announcements: Announcements and assignments will be posted via Moodle. Stu-
dents are expected to use their boun email addresses to get notifications.

Project: Project will be teamwork. Students are allowed to form teams of at most three
people.

Quiz: Quizzes might be online on Moodle or paper-based in class. Regardless, they will be
solved in the classroom. The quiz grade of students who were not present in the classroom but
still solved the online quiz will be zero.

Late Submission and Extensions: Late submissions will not be accepted. The project
report should be submitted to Moodle before the deadline, 23:59 on the due date. Submissions
sent over email after 23:59 will not be accepted, and the grade for late submission will be zero.
Personal extensions are not permitted. Please check the project report due dates beforehand
and plan accordingly.

Make-up exams: A single make-up exam will be given during the excuse exam period. This
make-up is for students whose application to the Excuse Committee (Mazeret Komisyonu) is
accepted and students who could not take the midterm due to health issues justified by a
doctor’s report. Therefore, there will NOT be two separate make-up exams. There will NOT
be any make-up quizzes.

Academic Integrity: In this class, assignments and exams should be completed individually.
Cheating of any sort is forbidden. The students involved in cheating will fail, and disciplinary
actions will be taken.

You might also like