| Instructor: | Chunyi Peng |
| Class hours/locations: | LWSN B134, MWF 10:30 am - 11:20 am |
| Office hours: | LWSN 2142E, F 13:00 - 14:00PM, or by appointment (send e-mail) |
Past readings (selected): Spring 2019, Fall 2017
This course is at the intersection of 5G and AI. Both are retransforming our society and life. 5G and beyond is re-defining connectivity, which is building a hyper-connected world with digitalizing everything from our physical world. AI, particularly ML, is giving new power to any computer system, including 5G and beyond, to run in a smarter way towards better performance, higher efficiency and enhanced merits desired by users.
This course is completely restructured to cover the state-of-art and emerging topics in the interaction of 5G and AI.
I offered its earlier versions in Spring 2019 (CS590: Network Intelligence in 5G and Beyond) and the first 5G course in Fall 2017 (CS590:Embracing 5G and Beyond 5G). Many thanks to constructive feedbacks from previous students in the class, and nice suggestions from professors, researchers and students in this area, I re-develop and re-structure this new course with a focus on AI for 5G and 5G for AI, along with thrilling breakthroughs in the recent years.
Two big changes: (1) The course lectures are organized with course lectures on basic and advanced topics towards network intelligence (like a reugular course), invited talk (like a seminar) and in-class paper presentation/discussion led by students. (2) This course focuses on cultivating skills from do-it-yourself practice. You will learn how US carrier networks (AT&T, Verizon, T-Mobile and Google Fi) work today, get a feel about what is good and what is not, understand why, and attempt to solve some problems on your own.
5G is on the way. Not only billions of users (like you) but also trillions of things (e.g., cars, drones, robots, wearables, sensors, actuators) will stay connected over the Internet. This not only further advance Mobile Internet to next level (everything is connected) but also reshape our life and society with unprecedented cyber-physical interactions like Metaverse, MixedVirtualAugmented Reality, smart citycommunityhome, autonomous driving, tactile internet, massive IoTs, remote healthcare, to name a few.
In this course, we are going to explore NETWORK INTELLIGENCE, one ongoing revolution to next-generation mobile network technology (5G) and beyond. Network intelligence is an inter-disciplined approach over data science, machine learning, distributed systems, programming language to develop and enhance system’s learning, reasoning and acting capabilities.
Throughout this course, students will learn key principles in mobile networking research, understand the state-of-art and recent trends, master a suite of research skills (e.g., paper reading, critique, critical thinking, problem solving, report writing, team work, communication, and presentation), and gain experience of carry out original research through course projects. Hopefully, through this course, students will generate publishable results from course projects or find some interesting topics for your long-term research.
In particular, you will learn and develop the following knowledge and skillset.
5G knowledge. You will systematically learn new demands, standard technologies and advanced topics for 5G and even beyond (say, 6G).
Network intelligence. You will learn this cutting-edge concepts and techniques when AI and data science are greatly reshaping our society. Here, you will learn how to conceptually adapt their principles and techniques in domain-specific problems to innovate our network access with learning-reasoning-acting capabilities in operation and also by design.
Skills to develop, enhance and exploit network intelligence. You will explore the topics of your interests and gain experience by carrying out original research projects.
No textbooks are required. The course materials are mainly from the lecturing slides I’ve made, the ongoing work from active researchers (invited talks) and recent research papers from top conferences like SIGCOMM, MOBICOM, NSDI, MobiSys, etc..
As an advanced topic course, we assume that students already have a basic understanding on networking fundamentals. The students must take CS422 (undergraduate) or CS536 (graduate) or equivalent courses before. Project experience and good programming skills are a must, as the course project is an important part of this class.
Homework and lab assignments: 30 %
In-class presentation, discussion and participation: 15%
Mid-term: 20%
Project: 35%
Homework and lab assignments: 3 homework/lab assignments on real-world networking problems.
In-class presentation: each student is going to present papers and participate paper discussion in class. Papers can be selected from the reading list or other top conferences (please let me know your pick in advance). It is encouraged to read the paper relevant to your project.
Midterm: An in-class midterm is to cover the lectures, homework and topics discussed as a regular course.
Course project: Each project will be done in a team of 2-4 students. Team and topic will be determined in the first few weeks. A course project report, with an in-class presentation in the final week is required.