EESC 7V86: Deep Learning in 5G Wireless (Spring 2025)

Instructor: Dr. Faris B. Mismar (fbm090020)

Meeting Days and Times: Monday and Wednesday at 5:30 pm to 6:45 pm Central Time.

Venue: FO 2.702

Office Hours: By Appointment.

Overview

Introduction to Python; review of linear algebra, probability, and statistics with Python; fundamentals of wireless communications; linear regression for channel modeling; classifiers in Python: symbol detection; deep learning in Python using Keras and TensorFlow; introduction to 5G New Radio (NR): NR overall architecture and protocol stack; NR physical, medium access control, and radio resource control layers; NR radio resource management procedures; fully connected feedforward neural networks; convolutional neural networks; recurrent neural networks; deep reinforcement learning; applications and use cases in 5G NR and beyond.

Topics

This class will cover the following topics:

• Part I: Intro to Python and Refreshers of Linear Algebra, Probability, and Statistics.

• Part II: Concepts on Machine Learning and Deep Learning.

• Part III: 5G Wireless Communications.

• Part IV: Use Cases of Deep Learning in 5G.

• Part V: Presentations.

Pre-requisites

Linear algebra, probability, and statistics. Familiarity with Python 3 and some of its relevant libraries (numpy, pandas, scikit-learn, and TensorFlow) will be helpful for the course project and assignments. We will review pre-requisites in class in the first two weeks.

Student Learning Objectives

This course is primarily designed for graduate students with background in wireless communications and practical research interest in deep learning and its applications to wireless communications. Upon completion of this course, students have the ability to:

• explain the 5G radio access network architecture, the 5G air interface protocol stack, and the various procedures at different layers in the stack

• explain the various deep learning algorithms and some use cases towards wireless procedures, and

• produce a written project report commensurate with guidelines and conventions prevailing in the industry.

Reading Material

F. B. Mismar, "A Quick Primer on Machine Learning in Wireless Communications", arXiv e-prints, arXiv:2312.17713. Available: https://arxiv.org/pdf/2312.17713. Code: GitHub.

Other material may be handed out from the suggested material below as needed.

Suggested Material

• Andreas F. Molisch, Wireless Communications, 2nd ed. West Sussex, UK: Wiley-IEEE Press, 2010. ISBN: 978-0521845274.

• Ian Goodfellow, Yoshua Bengio, and Aaron Courville, Deep Learning, 1st ed. Cambridge, MA, USA: The MIT Press, 2016. ISBN: 978-0262035613. Available: https://www.deeplearningbook.org/.

• Christopher M. Bishop, Pattern Recognition and Machine Learning, New York, NJ, USA: Springer, 2006. ISBN: 978-0387310732.

• Machine Learning For Communications Emerging Technologies Initiative https://mlc.committees.comsoc.org/.

• Open Radio Access Networks (O-RAN) Alliance. https://www.o-ran.org/.

• Third Generation Partnership Project (3gpp) RAN1 https://www.3gpp.org/Specifications-groups/ran-plenary/45-ran1-radio-layer-1/.

Grading

The class will be graded as follows:

Homework (20%), quizzes (10%), one-page project proposal (20%), and final project (50%). There are no exams. The final project is further broken down to: 15% oral in-class presentation and 35% term-paper due on the last day of instruction.

Part Scribing

One homework entails the task of scribing an entire part. A group of students volunteer to take detailed notes and upload them online in a readable format. All scribed notes will be assessed for quality and content. They are due no later than 11:59 pm Central US Time two days after we conclude a part. The instructor will provide a LaTeX template for scribing purposes.

At the end of the class, you will get a score out of 100 rounded up to two decimals and based on the percentages above. Letter grades: A 100-90%, A- 89.99-86.67%, B+ 86.66-83.34%, B 83.33-80.00%, etc.

Quizzes

Quizzes will be held in class for 15-30 mins. Their objective is to review key concepts introduced in class. There will be 6 pop quizzes and the least score will be dropped.

Final Project

Students will present their final project in the last two weeks of class. Details will be shared during the semester.