Deep Learning Algorithms and Methods
CSE 547_51
Instructor:
Office:
Contact:
Office Hrs:
Spring 2024
Dr. Hichem Frigui
204 Duthie Engineering Center
h.frigui@louisville.edu
T TR 1:00 PM — 1:55 PM, or by appointment
Course Description:
Deep Learning is an area of Artificial Intelligence and Machine Learning techniques. Its ability to
represent objects without the need of extensive feature engineering or pre-defining rules makes it
one of the most powerful and flexible frameworks to learn from data. This course covers basic
concepts and applications of Deep Learning Techniques.
Course Prerequisites:
Students are expected to have the following background:
 Basic knowledge of probability and statistics
 Basic understanding of college-level algebra and calculus.
 CSE 532, or IE 360 or equivalent
 Familiarity with Python is a plus, but early lectures and tutorials will address the fundamentals
required for general machine learning tasks.
Textbooks (References, not required):



Chollet, F. (2017). Deep learning with python. Manning Publications.
Stevens E, Antiga, L, Viehmann, T (2020) Deep Learning with PyTorch. Manning Publications.
Other online references will be posted on Blackboard regularly.
Course Format:





H. F.
Lectures will be recorded and posted on Blackboard. However, students are still required to attend
classes and participate.
Office hours: Tuesdays and Thursdays, 1:00pm – 1:55pm. You can also make an appointment for
another time slot.
Communication: The best way to reach me is through e-mail: h.frigui@louisville.edu
Homework: will be posted and submitted online.
Exam: Students will be given 24 hours to take the mid-term exam and submit the reports online.
Students will also be required to record and upload a 10 min presentation to explain their submitted
solution/analysis.
1 of 4
1/10/2024
Topics Covered









Introduction to traditional neural networks components
o Basics, Vector/Matrix Notation, Activation functions, Cost functions and gradient
descend optimization, Forward and Backward passes, Training, Testing, validation, and
over-fitting.
Python and Deep learning environments
o Brief introduction to Python, Data Representation (scalars, vectors and tensors),
Manipulating data with Python - Tensor operations, Deep Learning projects end-to-end,
Environments and Python libraries
Convolutional Neural Networks
o Convolutions, Activation functions, Pooling, Full connections, Optimization: Loss
functions; Gradient descent; mini-batch; batch normalization; Rule Dropout, Building,
training, and testing a CNN
Data Augmentation and Transfer Learning
o Pre-Trained Network Architectures
o AlexNet, GoogLeNet, ResNet, VGG, Others
Auto-Encoders (AE), Basics, Different types of AE, Building AE
Deep Recurrent Neural Networks
o Gated Recurrent Units (GRU)
o Long Short Term Memory (LSTM)
Word Embedding
Generative Adversarial Networks (GAN)
Applications
o Image and Text Classification
o Image and Text Generation
o Image Captioning
o Machine Translation
o Sentiment Analysis.
Objectives
1.
2.
3.
4.
Understand the fundamentals of Deep Learning Techniques.
Handle data (text, image, audio, among others) to be used in Machine Learning applications.
Understand basic mathematical concepts that support Deep Learning techniques.
Implement, train, and validate basic deep learning architectures and models using TensorFlow
and Keras API for Python
5. Demonstrate an ability to use appropriate deep learning architectures/models for different data
analysis tasks.
6. Understand some state-of-the-art Deep Neural Networks.
7. Demonstrate an ability to analyze and compare the performance of different DNN
architectures/models.
Evaluation
Students’ assessments consist of homework, a final project, written reports, and mid-term exam. All are
assigned and submitted on Blackboard.
 Homework (50%): several homework sets will be assigned. The purpose is to have every student
complete them with full understanding. Not all assignments are equally weighted, and all of them
H. F.
2 of 4
1/10/2024


would involve Python implementation and a written report. The report consists of a presentation
(e.g. PPT file) that needs to be (audio) recorded and submitted.
Project (30%): The final project consists of the design and implementation of a comprehensive
machine learning system that involves data from a real application. You will be provided with some
data to develop and train a machine learning system for a specific task. The goal is to learn a
solution that maximizes accuracy. The best performing system will be determined on some unseen
test data. The winner of this competition will be awarded additional bonus points to be added to
his/her final grade.The project grade will be based on effort, accuracy, presentation, discussion,
and analysis of the results.
Similar to the homework assignments, the project report is required to be a presentation (e.g.
PPT file) that needs to be (audio) recorded and submitted.
Mid-term exam (20%)
The format of the mid-term exam report is required to be a presentation (e.g. PPT file) that
needs to be (audio) recorded and submitted.
Grading scale:
[90, 100] = A
[80, 89] = B
[65, 79] = C
[50, 64] = D
[0, 49] = F
Incomplete
As described in the Academic Catalog (https://catalog.louisville.edu/undergraduate/universitywide-unit-specific-policies/incomplete-course-work/), students may request an incomplete under
the following conditions:
o More than 50% of the semester’s course work, as computed by percent of final grade
excluding the final exam has been completed.
o Performance on the completed course work is considered passing.
o The inability to complete the final portion of the course work is due to reasons beyond
the student’s control, such as, but not limited to: prolonged illness, death in the
immediate family, catastrophic events such as a car wreck, and travel restrictions.
Before a grade of incomplete is assigned, the student must provide documentation of the
extenuating circumstances. The faculty member must develop and provide the student with a
written statement outlining the schedule to complete the coursework. This must include dates and
times when all remaining assignments are to be completed; on what date any missing exams will
be taken, as well as the times and locations where exams will be proctored. Exams taken as part of
removing the incomplete can be different from the exams given during the semester. Both the
student and the instructor must agree in writing on the schedule. If the student does not meet the
deadlines in the fulfillment of the agreed upon schedule, a grade of F will be assigned, and no
additional extensions will be given.
Course Policies:


H. F.
Lecture slides, video recording of lectures, homework assignments, and other useful information will be
posted on the course web page.
You should carefully read the Speed School section on Academic Dishonesty Policy and Procedure (can be
found at http://louisville.edu/speed/academics/academicDishonesty )
3 of 4
1/10/2024
Computer Issues and IT Support:
Speed IT staff are available by appointment from 9 am to 4 pm to assist you with your technology needs. You may
schedule an appointment by sending a detailed email including any relevant error codes and screen snips at
SPDHelp@Louisville.edu (preferred) or 502-852-7620.
Title IX/Clery Act Notification
Sexual misconduct (including sexual harassment, sexual assault, and any other nonconsensual behavior of a sexual
nature) and sex discrimination violate University policies. Students experiencing such behavior may obtain
confidential support from the PEACC Program (852-2663), Counseling Center (852-6585), and Campus Health
Services (852-6479). To report sexual misconduct or sex discrimination, contact the Dean of Students (852-5787) or
University of Louisville Police (852-6111).
Disclosure to University faculty or instructors of sexual misconduct, domestic violence, dating violence, or sex
discrimination occurring on campus, in a University-sponsored program, or involving a campus visitor or University
student or employee (whether current or former) is not confidential under Title IX. Faculty and instructors must
forward such reports, including names and circumstances, to the University’s Title IX officer.
For more information, see http://louisville.edu/hr/employeerelations/sexual-misconduct-brochure.
Notes: The above material is subject to revision. Changes will be announced on Blackboard.
H. F.
4 of 4
1/10/2024