Thesis offers

This is precisely why companies such as Google have invested a lot of resources on exploring applications of GNN, some of which are as mediatic as Alpha Fold (recently solving the Protein-Folding problem), or used by Goolge Maps to make predictions of the expected traffic of a road. In this regard, the goal of this thesis is to continue investigating about the potential of GNNs when applied to the field of Computer Networks. Our goal is to provide GNN-based solutions that make computer networks run more autonomously and in a more efficient way.

Accurate prediction of these conditions for a sailing race is invaluable for sailors as it heavily influences their tactics during the race. In order to develop a decision support system applicable to the Olympic Classes Sailing Venues, the following components are being developed and integrated into a single web-based platform: Wind component, Waves component, Oceanic current component, Boat performance component.

The current master's thesis proposal is focused on the wind component.

The goal of the previous theses was to develop a methodology/procedure for analyzing wind datasets to automatically identify significant wind patterns. In other words, the aim was to recognize characteristic features of wind speed and direction related to other weather parameters of the day (such as air pressure, air and water temperature, cloud coverage, etc.) and the geographical position of the specific racing area. The analysis was conducted on both recorded wind data and wind data predicted by numerical prediction models. A highly promising correlation was found between the predicted and recorded data, which is believed to provide valuable insights for validating the weather model that will be used daily during the Olympic Games.

For instance, while traditional classification schemes for wind patterns are based on meteorological experience and manual analysis of synoptic weather charts, our current approach based on clustering can automatically induce wind patterns, including characteristic features, its evolution throughout the day, and specific characteristics related to the geographical position of the racing area. This allows for:

- A detailed analysis to determine the representativeness of the wind fields encountered during the measurement period, their frequency of occurrence, timing, rate of evolution, and transition probabilities.

- Consequently, a more accurate prediction of the expected conditions before a sailing race, which is highly valuable information for the sailors.

During the previous master projects, the clustering module was fully developed and applied to data from weather prediction models as well as data collected at sea. Traditional clustering techniques were employed for this purpose. Additionally, since the recorded data are collected daily and exhibit a time-series behavior, advanced and novel sequential clustering techniques were applied to extract more complex sequential wind patterns. Building on these results, the current project aims to expand and enhance the achievements thus far by:

1.  - Applying additional machine learning techniques to incorporate key parameters such as geolocation and weather parameters other than specific wind data.

2.  - Developing a prediction module that suggests the forecasted pattern and behavior of the wind for a racing day based on weather model data and, if available, actual data from the previous days and/or early on the current day.

3.  - Generalizing the environment so that it can be used not only in Marseille (the Olympic venue for Paris 2024) but also in any specific location where wind patterns are of interest and data is available.

The sea data is collected in the format provided by the Austrian Sailing Federation, which owns the on-board wind measurement system. The weather prediction models dataset follows the format used by operational meteorological centers, as suggested by the World Meteorological Organization, for the storage and exchange of gridded weather and climate model fields (grib2 or NetCDF).

Although this is a Modality A thesis, interaction and collaboration with TriM company are highly likely depending on the progress of the work.

Objectives:

- Understand the challenges of quantum communication for quantum computers.

- Develop new architectural methods to extend single-chip quantum computers to the multi-core paradigm.

- Model quantum interconnects within NetSquid, a simulator for quantum networks.

- Simulate different protocols for quantum communication within quantum computers.

We have developed LoRaMesher, an on-going implementation for doing mesh networking with LoRa nodes.

https://github.com/LoRaMesher/LoRaMesher

LoRaMesher is base on this protocol design: https://upcommons.upc.edu/bitstream/handle/2117/347444/A_minimalistic_distance_vector_routing_protocol_for_LoRa_mesh_networks.pdf?sequence=1

LoRa is a communication technology for IoT applications. It is typically used to send small amounts of data from the sensor node to a gateway (this is the so-called LoRaWAN architecture). But with LoRaMesher another network topology is also possible: to do mesh networking with LoRa among sensor nodes.

What for? E.g. for wildlife preservation:

https://www.hackster.io/mithun-das/mahout-save-the-elephants-819b54

Technical details about LoRaMesher: Code is in C++. The implementation leverages RTOS to perform the functions.

We deploy LoRaMesher on ESP32 LoRa boards, such as the T-Beam (e.g. https://tienda.bricogeek.com/arduino-compatibles/1502-ttgo-t-beam-esp32-wifi-gps-neo-6m-lora-868-mhz.html).  

A project with LoRaMesher can focus on a specific topic (embedded systems, network level, machine learning or application level) according to interest. 

* improving the LoRaMesher protocoll design and implementation, e.g.

- design, implementation and evaluation of different protocol alternatives for reliable messaging with LoRaMesher

- explore different metrics for the routing table in LoRaMesher (currently we use hops, but it is not reflecting the link quality, leading to suboptimal performance in heterogeneous links)

* machine learning.

- We have an initial federated learning prototype application that runs over LoRaMesher, a first steo to go further.

* applications that use LoRaMesher as a communication layer

- We have already developed a demo application called LoRaChat. A user connects the Android bluetooth terminal and can chat with other users in the LoRaMesh network.

https://github.com/Jaimi5/LoRaChat

A similar application is that of Meshtastic. https://meshtastic.org

- We currently work on an IoT monitoring application which spans over the LoRa mesh network and services in Internet. 

It could be very interesting to demonstrate this concept of interconnected LoRa and Internet services with a new application. We would like to create and test new applications which (in part) use the LoRaMesh network. These application can have an Android app (if the application is end user oriented) or the code can run entirely on the IoT board (maybe a "tracker" for a Peer-to-Peer system, and then integrate the LoRa mesh implementation as a library.

* embedded systems: We would like to port LoRaMesher to the Arduino Portenta board (LoRaMesher currently runs only on ESP32 boards). We have the board ready with LoRa radio. But the FreeRTOS usage that we make in LoRaMesher has to be "translated" into the equivalent in MBedOS.

The utilization of artificial intelligence (AI) for image-features recognition has been a theme of research studies in the last decades, among other yielding a number of publicly available databases containing pictures labelled with features of interest. One application that has gained significant interest is the recognition of human emotions, envisioned to be used among others in precision psychology and psychiatry. Several databases have been made publicly and freely available for the purpose of training AI models for emotion recognition [1], and such models have been demonstrated to feature rather promising emotion detection accuracy. However, this emotion detection is intrinsically privacy un-preserving, which limits its wider practical utilization.

In this project, we will aim at assessing the hypothesis that the same emotion recognition accuracy can be achieved when utilizing fine-grained 3D point-clouds of the human faces containing different emotions. As the first step in assessing this assumption, there is a need for creating 3D representations (i.e., point-clouds) of the human faces utilizing 2D images of the spatial expressions from the publicly available databases. The student is expected to bridge that gap, i.e., to design and implement an AI-tool that takes as an input a 2D image of a human facial expression and outputs its fine-grained 3D representation. The main development tool for the project is Python and its libraries.

[1] Available online: https://en.wikipedia.org/wiki/List_of_facial_expression_databases

At the one hand, virtual reality (VR) experiences are becoming increasingly more realistic, primarily due to novel technological advances in supporting both rotational and translational mobility of VR users, as well as enabling multiuser experiences [1]. On the other, older generations of VR technology supporting solely rotational mobility of the VR users are experiencing rapid deployments in appropriate physical VR setups, where the users are increasingly gaining the opportunity to dynamically switch between different types of experiences. In the near future, such deployments can be forecast for multiuser full-immersive VR technology encompassing support for both rotational and translational movements. Such support is expected to be based on redirected walking (RDW), a concept of imperceivable re-steering of mobile VR users within a constrained physical space.

The feasibility of RDW in imperceivably steering multiple users within the same physical space has been demonstrated under the assumption of the number and mobility patterns of the users being kept constant for the full duration of a full-immersive multiuser VR experience. However, based on the experiences with older generations of VR technology, it is to be expected that the users will dynamical enter an exit a shared physical space, in which they will also feature the possibility of independently switching between different types of experiences, potentially resulting in sudden changes in their physical mobility patterns. The main goal of this project is to assess the effects of dynamic changes in the number of co-existing VR users, as well changes in the mobility patterns of the users in the physical setups. Based on the assessed effects (and in case of a longer project such as BSc/MSc thesis), the student is envisioned to propose a method for appropriate scaling of the number of users based on their mobility patterns, sizes of deployment environments, obstacles in the deployment environment (e.g., other users). The student is expected to gain experiences with next generation VR technology through this primarily simulation-based study using an existing Python-based simulator for redirected walking [2].

[1] Van Onsem, T., Struye, J., Costa-Pérez, X., Famaey, J., Lemic, F. "Toward Full-immersive Multiuser Virtual Reality with Redirected Walking", IEEE Access, 2023

[2] Lemic, F., Struye, J., Famaey J. "User Mobility Simulator for Full-Immersive Multiuser Virtual Reality with Redirected Walking", 12th ACM Multimedia Systems Conference (MMSys), 2021

Cybersecurity is becoming an increasingly important challenge for all companies and individuals alike. While big names used to be the main targets in the past, as people's lives move online, anyone is nowadays a potential target for any kind of cyber-attack, ranging from phishing to ransomware or serious privacy issues. In order to fight against those ever-evolving threats, Machine Learning is increasingly being used behind the scenes to design better systems that are able of self-learning to boost detection rates and boost overall resilience to unknown attacks. As AI-based solutions penetrate products across the industry, a new kind of threat that is often overlooked is becoming more and more prominent and dangerous: adversarial machine learning (AML).

AML focuses on designing specific inputs to deceive a previously trained Machine Learning models into misclassifying them for a specific purpose. One of the main flaws of any state-of-the-art Machine Learning or Deep Learning algorithms is that they assume that the nature of the data they receive is systematically benign, which is generally the case but does not hold true when an adversarial input is received. The motivation behind altering a ML model into thinking that, for example, a new sample is benign when in fact is malicious can range from pure research to more serious real-life issues such as an autonomous car wrongly classifying a stop sign (and thus provoking a fatal accident) or a wrongly diagnosed disease because of a slightly manipulated magnetic resonance image.

This problem is no exception for Cybersecurity where companies wrongly assume that once the last AI-based product is deployed in their network, their employees are safe...

Mobile networks are continuously experiencing a profound transformation, driven by the emergence of groundbreaking applications and communication paradigms, such as the eagerly expected metaverse. In the coming years networks will face daunting challenges to meet the demands of these future applications. In this (r)evolution, Deep Learning is among the most promising techniques for achieving unprecedented levels of connectivity and user experience that can meet such demands.

In this project, you will have the opportunity to work on AI and machine learning applied to real-world problems around 5G and beyond networks. We will focus on hard open challenges, such as AI-based mobility management optimization and planning in future 6G networks, or creating cutting-edge network automation technology to facilitate the day-to-day operations of network engineers around the world.

 The goal of this thesis is to enable the application of Deep Reinforcement Learning techniques for optimization problems in very large and complex scenarios, in this case Data Center Networks. Inspired by the last related work from OpenAI (https://openai.com/blog/evolution-strategies/), we will explore time-efficient alternatives to train a Graph Neural Network based DRL agent on large optimization scenarios in a reasonable amount of time.

In many teaching applications such as the study of anatomy, the student is required to obtain a three-dimensional view of the models or structures to be analyzed. This goal is necessary for both individual study and class follow-up. There are some VR-based systems that usually use immersive VR HMD. However, these systems require an infrastructure and cost that is often not affordable.

This master's thesis aims to analyze the feasibility of a remote VR system based on the use of mobile devices with cardboard glasses and low-cost interaction devices. It will start from a system based on HTC-VIVES programmed with Unity. Different portability alternatives to the new platform will be analyzed both in terms of the rendering of the models (locally or on a server) and the limitations of the interaction and connection between students and teacher. Once the advantages/limitations have been analyzed, a prototype will be developed with basic interaction techniques and its usability will be analyzed. The models that will be visualized will be anatomical structures modeled with triangles. For more information contact the directors.

Internet services are known to collect large amounts of personal information. As a result, more than half of EU citizens are concerned about their online privacy. In this context, data brokers are companies devoted to collecting and selling personal information to other companies. Data brokers are often implemented as third-party trackers, which allow them to gain visibility across the Internet. Recent works show this personal information is not only used for targeted advertising, but also for more obscure practices, such as price discrimination, credit scoring, phishing or identity theft. The project addresses these growing privacy concerns by enhancing ePrivo, a new online privacy observatory. ePrivo will continuously scan the Internet to unveil third-party trackers, the tracking methods they use and the data brokers behind them. A proof-of-concept prototype of ePrivo is already available at https://eprivo.eu. In this project, we will deploy the ePrivo service in production, extend its capabilities for web tracking and phishing detection, and release it under an open-source license. The results will be useful to the RIPE community, including ISPs, network operators, Internet users, policy makers and researchers.

TinyML is a growing community that does machine learning on microcontrollers https://www.tinyml.org. 

Microcontrollers are sometimes the only choice when the power supply is limited, e.g. in for battery- or solar-powered applications. Application examples of TinyML are "wildlife" observation, such as: https://mybirdbuddy.com/ https://opencollar.io/, but tiny machine learning is also in domestic, healthcare and industrial applications.

While Arduino Uno boards are well known for all kinds of hobbyist microcontroller projects, for machine learning the more powerful 32 bit microcontrollers and development boards are used, such as the Arduino Portenta H7 or Arduino Nano 33 BLE Sense. These are able to run machine learning applications. To get the first information on this topic have a look at TensorFlow Lite for Microcontrollers: https://www.tensorflow.org/lite/microcontrollers.

In this project we would like to explore on-device machine learning model training specifically doing it by federated learning. We would like to use the Arduino Portenta H7, the most powerful microcontroller board from the Arduino family. We have Arduino Portenta H7 that can be used for the project. 

We have done previous work and code available where we showed that it is feasible.

https://www.mdpi.com/2079-9292/11/4/573

In the project it would be interesting to explore the capacities of the Portant H7, like using the two cores of the board, use at least one the sensors (camera, accelaration, microphone,...), and use the network connectivity (Wifi, BLE, LoRa), try with Convolutionary Neural Networks, ...

One concrete case would be to use for the communication part of federated learning the Wifi connectivity of the Arduino Portenta H7 to build a "cluster" (inspired by clusters of Raspberry Pi) where all the nodes are interconnected and can participate in federated learning. 

The project can start with already working code (in C/C++, Python) that we have from previous work on TinyML with on-device training and federated learning. https://upcommons.upc.edu/bitstream/handle/2117/353756/160036.pdf

This proposal explores a specific case of the simulation of weathering effects in computer graphics. Weathering refers to the process of deterioration and transformation that occurs on surfaces over time due to exposure to various environmental factors. In this project, the main focus will be on simulating and visualizing the growth of lichens on cultural heritage 3D models. Lichens are unique organisms that colonize surfaces and contribute to changes in the appearance and degradation of materials. This project aims to accurately model the growth patterns of lichens on the surface of the models and render the changes in the appearance of the materials in terms of color and/or shape. The project will involve implementing algorithms that simulate the growth of lichens based on environmental conditions and surface characteristics. The final outcome will be a visually compelling and scientifically accurate representation of the weathering process caused by lichens on cultural heritage 3D models, contributing to the preservation and conservation of these invaluable artifacts.