Quantum Federated Learning in Distributed Systems: Challenges and Future Directions

This project reviews quantum federated learning (QFL) for non-IID and heterogeneous distributed systems. Practical federated learning systems often involve clients with different data distributions, device capabilities, and communication conditions, which can lead to unstable aggregation and degraded global performance. QFL provides a promising direction by combining federated learning with quantum neural networks and parameterized quantum […]

Read More
Machine Learning-Based Assessment of Masonry Degradation in Cold and Hot Climates

Heritage masonry buildings are deteriorating at an accelerating rate due to aging, weathering, and environmental exposure. Early and accurate detection of surface defects such as cracks, mortar loss, and spalling is essential to prevent progressive structural damage and enable timely, cost-effective conservation. However, current practice relies predominantly on manual visual inspection, which is subjective, labor-intensive, […]

Read More
Designing Responsible and Efficient AI Systems for IoT Threat Detection and Retrieval-Augmented Educational Assistance

This project explores two critical areas of AI research: cybersecurity for Internet of Things (IoT) systems and responsible AI assistants for education and research. IoT devices, widely used in healthcare, smart cities, and industry, are highly vulnerable to cyberattacks due to limited resources and large-scale connectivity. We propose AI-based intrusion detection systems that are lightweight, […]

Read More
STARTUP DEVELOPEMENT IN THE ERA OF GENERATIVE AI

The proposed project investigates how Generative Artificial Intelligence (GenAI) is changing the way startups recognize, evaluate, and develop new business opportunities. A visiting PhD student from IE University will work with Dr. Serghei Musaji at the University of New Brunswick to study Canadian startups and their local ecosystem, collecting original data through surveys, interviews, and […]

Read More
Low-cost LLM inference at the Edge

Seamlessly serving Large Language Model (LLM ) inference requests/queries generated from user devices such as smartphones, laptops, tablets, etc., is critical for LLM service providers. The current approach of serving these queries using large LLMs in the cloud is expensive and incurs high latency and energy costs. More importantly, it cannot meet the growing demand […]

Read More
Detection of Calving Activity in Dairy Cows Through Computer Vision

Calving is a critical event in dairy farming, as abnormal calving can lead to calf mortality, health complications for the cow, and economic losses for farmers. These risks can be minimized by the timely detection of calving onset, which can be determined from animal behaviour. Ear position is a behavioural indicator of cows’ emotional states […]

Read More
A Multi-Modal Vision Framework for Non-Destructive Evaluation of Thermal Spray Coatings

Thermal spray coatings are thin protective layers used on aircraft engines, turbines, and other high-temperature equipment to prevent damage from heat and wear. Today, checking the quality of these coatings is slow, expensive, and often destructive: samples must be cut, polished, and tested using several separate techniques. This project aims to make coating inspection faster, […]

Read More
Outlier Generation in Diffusion Models

The growing prevalence of anomalous events in real-world data highlights the importance of synthetic data generation techniques capable of identifying and generating outliers. This research introduces a novel approach for generating controlled outliers in score-based generative models by modifying diffusion model dynamics. We propose incorporating a drift control term to the reverse stochastic differential equation […]

Read More
Bentley Systems: Spécialiser un modèle de segmentation pour extraire les labels et les lignes de repères des plans d’ingénierie et reproduire l’esthétique d’un placement humain

1) L’idée principale du stage est de pouvoir prendre des dessins finaux (au format .pdf ou image) et à l’aide d’un modèle de segmentation, d’en retirer automatiquement les étiquettes et lignes de repères pour retrouver un dessin original non annoté et un fichier de définition de toutes les étiquettes incluant leur contenu en texte, leur […]

Read More
Ontology-based Host Similarity for Predicting Virus Host Range

This project explores how the biological relatedness of species—measured using the official “tree of life” taxonomy—can help predict which animals or plants a virus might infect. The student will build a similarity map between organisms using public databases, then test whether viruses that infect one species are likely to infect its close relatives. This research […]

Read More
Automatic identification of metaparameters in machine learning and signal processing applications (with UNaM)

This project will focus on the methods and algorithms which facilitate automatic choice of the parameters required in estimation/learning from data. We will focus on real-time algorithms such as those appearing in adaptive signal processing and on-line learning. The project will be particularly focus on the identification of meta-parameters in real-time from the data while […]

Read More
Détection d’épisodes d’hypoglycémie à travers des signaux ECG par d’IA et approche quantique.

Les épisodes d’hypoglycémie, des baisses dangereuses du taux de sucre dans le sang, constituent un risque majeur pour les personnes atteintes de diabète. Non détectés à temps, ils peuvent entraîner des complications graves comme des pertes de conscience ou des convulsions. Les dispositifs actuels de mesure continue du glucose (CGM), bien que utiles, restent coûteux, […]

Read More