Master Seminar (SoSe2020): Applications of Machine and Deep Learning in Mobile Networking
IMPORTANT: Due to the ongoing public health crisis, the seminar will be completely virtual. The first lecture will be published online during the first week of the summer semester (week 17). Please refer to the section "Course organization" for the timeline of course activities!
In the recent years, theoretical and technical advancements in the area of machine/deep learning, fuelled by the explosion of big data and computing resources, have attracted interests of practitioners in different fields and research areas. At the same time, (mobile) networking systems have evolved into complex, heterogeneous environments thanks to the exponentially increasing mobile traffic volumes, application of agile management of network resources in order to maximize user experience, and the requirement to support fine-grained real-time analytics. To tackle these problems, and to provide solutions that will aid managing evolving mobile systems, networking researchers are utilizing advanced machine learning techniques.
In this seminar course, we will study machine leaning approaches to solving mobile networking—and mobile applications—related research problems. The papers addressing these problems can be broadly classified into the following areas:
- Mobile Data Analysis
- Mobility Analysis and User Localization
- Network Security
- Network Control
- Mobile Network Applications
We will provide an initial list of topics and papers associated with those topics. The topic/paper choices are, however, not limited to the proposed topics, and everyone’s own ideas—falling into the scope of the seminar—are more than welcome.
Learning outcomes (study goals)
The seminar is envisioned to provide students with a review-based insight into practical applications of machine learning in the mobile networking domain. Upon completion of the seminar, the students will not only have solidified their understanding of certain machine learning algorithms and areas in networking research, but will also have became familiar with reading, reviewing and presenting academic papers in a setting similar to a conference or a workshop.
(Recommended) course requirements
- The participants are expected to have taken an undergraduate-level course on machine learning.
- The topics require also background in communications and networking technologies.
- Bachelor's degree in computer science or a related field is required.
- Ability to write and present in English.
Teaching and learning methods
Through extensive group discussions and individual student presentations we will dissect machine learning algorithms presented in the reviewed papers. Each student will present two papers, and review two additional papers, via the form that will be provided at the beginning of the course. All submissions and presentations are expected to be given in English.
The grading scheme is the following:
- Two written paper reviews (40%)
- Presentations of two papers (~20 minutes each) (50%)
- Active participation in the course: attending seminars and taking part in general discussion (10%)
Registration for the course
Registration is done using the Matching System of the department: www.in.tum.de/en/current-students/modules-and-courses/practical-courses-and-seminar-courses.html (you have to use the matching system to participate in the seminar!)
- Pre-meeting - 31.01.2020 at 9:15 in room 01.07.023
- Slides available here
- First lecture (online) - 22.04.2020, will be published in Moodle
- Informal meeting (virtual) - 24.04.2020, 11:00-11:30 link
- Topic selection deadline - 01.05.2020
- Topic allocation and scheduling - 04.05.2020
- Weekly presentations (virtual) - Fridays 16:00-18:00, first session 15.05.2020 link