Modern Mobile Communication Systems are growing in complexity due to the proliferation of 6G use cases, network slicing, and GenAI/AI. Operators must simultaneously meet diverse Service Level Agreements (SLAs) across slices (e.g. enhanced mobile broadband vs. ultra-reliable low-latency), optimize spectrum and power, and handle dynamic interference and traffic variations. Traditional approaches to network management rely on static configurations and specialized algorithms for each task (e.g. separate functions for handover optimization, slicing, interference mitigation). Even emerging network architectures (O-RAN, AI-RAN) use multiple specialized applications (e.g. xApp) for specific optimization tasks, leading to siloed solutions that lack a unified intelligence. This fragmented approach struggles to coordinate conflicting objectives (for instance, when different network slices compete for shared resources) and cannot easily adapt to unforeseen scenarios.
In this course, we will study the next generation mobile network from the architectural perspective and its evolution from legacy networks toward autonomous networks with built-in network automation and intelligence. In particular, we will review how telecommunication networks are evolved with the emergence of GenAI/AI (e.g. DeepSeek and OpenAI) and cloud-native computing (e.g. Kubernetes) to significantly increase the efficacy and agility of network management and operations allowing operators to dynamically optimize networks based on the high-level objectives (e.g. reduce energy consumption). In this regard, we will design, build and deploy autonomous agents powered by Large Language Models (LLMs) for real-time decision-making on a live cloud-native 5G network. We will also discuss how future autonomous networks can self-synthesize themselves to compose new capabilities and knowledge and as a result evolve over time and space. Finally, we review real world scenarios in both public and private telecommunication networks and how they could benefit from autonomous networking. The course is a mix of theory and lab sessions.
Learning outcomes:
- Be able to flexibly design a 5G network architecture.
- Be able to automate 5G network deployment.
- Be able to develop 5G network intelligence in a form of O-RAN xApp/rApps.
The course is roughly divided in two separate parts. The first covers the topics of computer forensics and incident response. In particular, we discuss a number of techniques and open source tools to acquire and analyze network traces, hard disk images, Windows and Linux operating system artifacts, log files, and memory images.
The second part of the course deals with the analysis of malware and unknown binaries. Here the goal is to introduce students to the main classes of techniques used in malware analysis and reverse engineering. We cover both static techniques (ELF and PE file structures, dissasseblers and decompilers, data and control flow analysis, abstract interpretation, ...) and dynamic techniques (sandboxing, library and syscall traces, dynamic instrumentation, debugging, taint analysis, unpacking,...). We will use mostly open source tools, with the exception of IDA Pro.
Teaching and Learning Methods : Lectures and Homework Assignment
Teaching and Learning Methods: Lectures, Homework (2-3), and a case study (study and present a research paper in a group of 2-3 students).
Course Policies: Mandatory participation, Case study optional, but recommended.
Scope
This course builds on the contents of Digicom1 and adds three new lab sessions covering the coding formats used in 3GPP and their corresponding receiver structures and decoding algorithms. The first part deals with receivers for bit-interleaved coded-modulation (BICM) for multi-layer transmission and a corresponding lab session on receivers for BICM with a single-layer. The second component provides an introduction to polar coding and decoding with a corresponding lab session on the 3GPP polar decoder for the physical downlink broadcast channel. The third component deals with low-density parity-check coding (LDPC) and includes a lab-session on the 3GPP LDPC decoder for the downlink shared channel (DLSCH).
Outline
- Bit-interleaved coded modulation for multi-layer transmission. Lab session on single-layer BICM reception of QAM signals
- Polar coding and the 3GPP Polar Code. Lab session on detection of the 3GPP PBCH (physical broadcast channel)
- LDPC coding and the 3GPP LDPC code. Lab session on detection of the 3GPP PDSCH/DLSCH.
Organization
- Integrated Lecture / Labs
- Autonomous labwork
Human-computer interaction (HCI) is the study of interaction between people (users) and computers, as the intersection of computer science, behavioral sciences, design and several other fields of study. This course aims to provide the basic concepts of user centered design when developing web applications. It will offer a deep dive presentation of modern web technologies: HTML-5, CSS-3 and Javascript. Finally, this course will provide techniques for evaluating user interfaces.
Teaching and Learning Methods: Lectures and mini-project (group of 3-4 students)
Course Policies: Attendance to course session is mandatory.
Course Description:
- Learn and understand the role of sketching in design and its relation to creativity
- Focus on interaction design for web applications
- Get an overview of existing software tools that support sketching and learn Sketchify or Balsamiq
- Practice and master new web technologies
- Develop rich web interactive applications (HTML5, CSS3, javascript)
- Learn advanced HTML5 APIs (web audio, media fragments, SEO and semantics)
- Be aware of accessibility constraints and devices constraints (mobile phone)
- Know how to evaluate user interfaces
- Know your users through user centered design methods
- Learn how to perform A/B tests, usability and ergonomics studies
Bibliography:
- Bill Buxton (2007), Sketching User Experiences: Getting the Design Right and the Right Design, Morgan Kaufmann.
- Marti A. Hearst (2009). Search User Interfaces, Cambridge. http://www.searchuserinterfaces.com/
- Review of User Interface Prototyping Tools, http://www.dexodesign.com/2008/11/07/review-16-user-interface-prototyping-tools/
- HTML EdX online course: https://www.edx.org/xseries/html5-w3c
Requirements:
- Software development methodologies
- Basic knowledge of web technologies (html, css, javscript) is a plus but not mandatory
Wireless communications are pervasive and have been used for a century. They are used in a very large set of security applications (communications by security forces, car key remote, alarm system, access control, drone command and control, surveillance devices) . However, day to day applications also require to be protected for privacy and personal security, such as WiFi or mobile communications (2G/3G/4G). At the same a number of challenges are present in wireless communications security, for example, messages are broadcasted, making it possible to intercept them without being noticed. Wireless signals are subject to jamming, making them unavailable.
This course will give a large perspective of the fundamental challenges in securing wireless communications, from the physical layer, modulations to the application protocols. A special focus will be put on practice with hands on exercises (using software defined radios and WiFi dongles).
Teaching and Learning Methods : Course is composed of lectures, Labs and small projects with final presentation.
Course Policies : Class attendance, labs and projects mandatory.
This module addresses mechanisms and strategies to improve traffic efficiency and carbon footprint. The objectives are first to introduce the underlying theory such as Waldrop Equilibrium required for efficient path planning. Second, it describes concepts and theory behind the optimization of traffic lights to traffic conditions. Third, it provides guidelines and methodologies to model emissions and integrate them into efficient path planning. Finally, it trains on best practices to apply these concepts for efficient and green path planning on vehicular traffic simulators.
Teaching and Learning Methods : Lectures and Lab sessions (group of 2 students)
Course Policies : Attendance to Lab session is mandatory
This module addresses mechanisms and strategies to model multi-modal transportation. The objectives are first to introduce concepts of population modeling. Second, it extends graph theory concepts for modeling heterogeneous transportation networks, in particular public transport networks. Mechanisms for activity and demand modeling for multimodal transportation will also be discussed. Finally, this lecture trains on best practices to apply these concepts for efficient multimodal transportation planning on vehicular traffic simulator
Teaching and Learning Methods : Lectures and Lab sessions (group of 2 students)
Course Policies : Attendance to Lab session is mandatory.
This course covers the Low Power Wide Area Network (LPWAN) protocols dedicated to IOT. LPWAN is a technology that intends to offer Internet connectivity to a large number of objects ("Things") under very strict requirements in terms of cost, power consumption, long distance, battery life, indoor penetration, etc. This course presents two families of LPWA protocols specially developed for IOT: (i) LPWAN for unlicensed spectrum (e.g. LoRa, SigFox ...) and (ii) Cellular LPWAN for licensed spectrum (e.g. 3GPP LTE Cat. M1 or Cat. NB).
Teaching and Learning Methods: The course is organized in 4 lectures and 3 labs.
Course Policies: Labs are Mandatory (reports)
This course covers the application-level protocols dedicated to IOT. Knowing the limited capacity, in terms of battery and CPU, of the things, the classical application protocols used in the Internet like HTTP are not adequate. This course presents the recent application protocols specially developed for IOT. These protocols are organized into two categories: (i) Client/server (like COAP) and (ii) Publish/Subscribe (like MQTT, XMPP). In addition to these protocols, this course introduces two types of architecture, specifically dedicated to host IOT services, like 3GPP MTC and oneM2M.
Teaching and Learning Methods: The course is organised in lectures and labs.
Course Policies: Labs are Mandatory (reports)
This course focuses on the principles of learning from data and quantification of uncertainty in the context of machine learning, by complementing and enriching the “Machine Learning and Intelligence Systems” course. The presentation of the material follows a common thread based on the probabilistic data modeling approach, so that many classical learning models/algorithms can be seen as special cases of inference problems for more general probabilistic models.
We will start by drawing connections between loss optimization and probabilistic inference (e.g., maximum likelihood estimation, maximum a posteriori estimation). We will then introduce the concept of Bayesian inference, and discuss how to perform inference in complex and intractable models using approximate methods (e.g., variational inference, Markov Chain Monte Carlo, Laplace approximation). We will then focus on prediction, and the evaluation of predictive models. We will start by discussing simple predictive models (e.g., linear regression, logistic regression), and then move on to more complex models (e.g., Gaussian processes, neural networks). We will discuss all these models in the context of probabilistic inference, and we will put in practice the probabilistic methods introduced in the first half. While mostly focused on supervised learning, we will also take a look at unsupervised learning and probabilistic generative models. Finally, the course will be complemented by several practical sessions, where students will be able to implement and experiment with the methods discussed in class (using Python).
Teaching and Learning Methods : Lectures and Lab sessions (preferably one student per group)
Course Policies : Attendance of Lab sessions is mandatory
This course provides an introduction to the automatic processing of speech and audio signals. It starts with a treatment of the human speech production and perception mechanisms and looks at how our understanding of them has influenced attempts to process speech and audio signals automatically. The course then considers the analysis, coding and parameterisation of signals in the case of different speech and audio processing tasks. After an introduction to essential pattern recognition techniques, the course considers specific applications including speech recognition, speaker recognition and speaker diarization. The course also includes a treatment of speech and audio coding, noise compensation and speech enhancement.
Teaching and Learning Methods:
The course is comprised of lectures and exercises and laboratory sessions.
Course policies: Attendance of laboratory sessions is mandatory.
<p>
Image & Video processing is part of many applications related to security: digital watermarking, steganography, image forensics, biometrics, and video surveillance.</p>
<ul>
<li>
<strong>Digital Watermarking </strong>allows owners or providers to hide an invisible and robust message inside a digital Multimedia document, mainly for security purposes such as owner or content authentication. There is a complex trade-off between the different parameters : capacity, visibility and robustness.</li>
<li>
<strong>Steganography</strong>is the art and science of writing hidden messages (in a picture or a video) in such a way that no-one apart from the sender and intended recipient even realizes there <em>is</em> a hidden message.</li>
<li>
<strong>Image Forensics </strong>includes two main objectives: (1) To determine through which data acquisition device a given image is generated; (2) To determine whether a given image has undergone any form of modification or processing.</li>
<li>
<strong>Biometrics</strong>: The security fields uses three different types of authentication : something you know, something you have, ore something you are : a biometric. Common physical biometrics includes fingerprints, hand geometry ; and retina, iris or facial characteristics. Behavioural characters include signature, voice. Ultimately, the technologies could find their strongest role as intertwined and complementary pieces of a multifactor authentication system. In the future biometrics is seen playing a key role in enhancing security, residing in smart cards and supporting personalized Web e-commerce services. Personalization through person authentication is also very appealing in the consumer product area. This course will focus on enabling technologies for Biometrics, with a particular emphasis on person verification and authentication based on or widely using image/video processing.</li>
<li>
<strong>Video surveillance </strong>is the monitoring of the behavior, activities, or other changing information, usually of people for the purpose of influencing, managing, directing, or protecting. By default, for a better scene understanding, automatic image processing tools are used between acquisition/transmission and visualization or storage</li>
</ul>
</div>
This module addresses the access methods in Wireless Local Access Networks (WLAN). The basic contention and management mechanisms are detailed. Current and emerging standards of WLAN toward 5G are also presented.
Teaching and Learning Methods : Lectures and Lab sessions (group of 2 students)
Course Policies : Attendance to Lab session is mandatory.