ECE 69500 - Machine Learning-Driven Computer Networking

Course Details

Lecture Hours: 3 Credits: 3

Areas of Specialization:

  • Computer Engineering

Counts as:

Normally Offered:

Each Fall

Campus/Online:

On-campus only

Requisites:

ECE 46300 or ECE 50863 or equivalent or graduate standing

Requisites by Topic:

Undergraduate course in computer networking

Catalog Description:

This is a graduate level class on computer networking and machine learning (ML). We will explore the state of the art in how ML and AI are being used in computer Networking. A key theme is off-the-shelf ML is often not sufficient, and there is need for new techniques at the intersection of networking and ML. Application areas include video streaming, congestion control, routing, network diagnosis, network security, and network configuration. We will also explore the state-of-the-art in how modern networks are themselves being rearchitected to enable futuristic needs of AI and ML applications. The course will involve lectures by the instructor and paper presentation and discussion with students. Students will conduct a semester long project at the intersection of ML and networking.

Required Text(s):

None.

Recommended Text(s):

None.

Lecture Outline:

Major Topics
1 Introduction and overview - 1 week
2 ML-Driven Routing - 2 weeks
3 ML for Internet Video - 2 weeks
4 Causal Reasoning and what-if questions - 1 week
5 ML for Network Diagnosis - 1 week
6 Intent-based Networking and LLMs - 1 week
7 Network Security and Anomaly Detection - 2 weeks
8 Network support for Distributed ML training - 3 weeks
9 Mid Term Exam and In Class Presentations - 2 weeks

Assessment Method:

Projects, presentations, and exams.