Projets novateurs réalisés

Subsurface biogeochemistry of hydrogen (H2) transformations: microbial implications for short-term energy storage

Hydrogen (H2) offers a strategic opportunity for Canada’s energy transition and more broadly is viewed as an essential part of the global solution to reducing carbon emissions and addressing climate change. Industry, governments and research agencies at regional and national scales are looking at the feasibility of H2 hubs that incorporate local logistics for H2 generation, storage and end-use. In this context, storage of H2 at scale will be a critical part in establishing H2 economies and markets. The scale of storage that is needed is realistically only available in subsurface settings, with salt caverns being considered a leading option. Since H2 is an excellent food (energy) source for microorganisms living in subsurface habitats, it is important to assess the potential interactions between “deep biosphere” microbiomes and stored H2 to understand the potential for H2 losses due to biological activity. Genomics information will lead to better predictions about the fate of stored H2 and whether strategies for mitigating microbial activity are needed.

Enhancing the Accuracy and Interpretability of Canadian Macro-Financial Tail Risk Forecasts via Multi-Quantile Deep Learning with Feature Engineering to Monitor Systemic Risks at the Bank of Canada

Crises risks are notoriously hard to quantify. Yet, when systemic crises materialize, for instance the Global Financial Crisis (GFC), the cost for the economy and the society can be huge, with protracted recessions and financial hardships for firms and households. Thus, it is essential for public authorities to monitor and proactively address systemic risks, thereby ensuring a stable and efficient financial system that can sustain economic growth and raise standards of living.

In this context, the Bank of Canada seeks to leverage advanced tools such as artificial intelligence and machine learning to keep improving its assessment of systemic risks. The purpose of this joint project with the academic partners is to develop state-of-the-art forecasts of macro-financial tail risks, capturing extreme shocks like those seen during the GFC.

The development of self-powered wearable biosensors

The proposed project aims to develop a wearable biosensor for real-time monitoring of critical cardiovascular disease (CVD) biomarkers, such as C-reactive protein, Troponin I, and Myoglobin, along with physiological parameters like heart rate. The biosensor will achieve high selectivity and durability by utilizing molecularly imprinted polymer (MIP) technology. The intern will collaborate with Professor Joseph Wang to integrate these biosensors into wearable devices, enabling continuous monitoring and early diagnosis of CVD. This project will benefit participating institutions by advancing wearable health technology, improving early CVD detection, and contributing to digital healthcare innovations.

ESROP – Osaka – Systems Optimization and Decision Making

This project will investigate new methods for improving decision-making tools that help organizations make better choices when faced with uncertainty. Specifically, we will study how to more accurately estimate weights in the Analytic Hierarchy Process (AHP), a common tool used for Multi-Criteria Decision Making (MCDM). By focusing on interval and fuzzy weight estimation, the project aims to create more reliable and efficient decision-support systems. The results will benefit participating institutions by providing better tools for complex decision-making, helping them prioritize actions and allocate resources more effectively in areas such as environmental management, policy planning, and other fields that require careful evaluation of multiple factors.

Segmented Seal Project

This research seeks to improve gas turbine engine components, focusing on dynamic seals that withstand extreme conditions, such as high temperatures, pressures, and abrasive particles, to enhance aircraft performance and durability. Current carbon-based materials used in seals like segmented seals—designed to prevent oil leakage—are vulnerable to degradation, which can lead to engine damage and costly maintenance. Although progress has been made in understanding their mechanical properties, there is limited knowledge on how these materials perform tribologically in realistic engine environments. This study aims to bridge this gap by employing a Category IV test rig for comprehensive testing, more closely simulating the harsh conditions seals experience in engines.

3D Point Cloud Foundation Model Project

Professor Kim’s team is advancing AIST’s initiative to develop a Foundation Model for Computer Vision, a transformative AI system designed to address diverse tasks through large-scale pre-training. These models are vital for computer vision, which focuses on enabling machines to interpret visual data like images and 3D point clouds. Professor Kim’s work centers on 3D point clouds, which are critical for applications such as forest management, urban planning, and terrain analysis. I will contribute my extensive expertise in 3D point cloud processing—supported by multiple publications on 3D semantic segmentation and digital twin development—to enhance the team’s ability to develop deep learning models for tasks such as tree species classification, terrain scene recognition, and point cloud semantic segmentation.

This collaboration will deliver valuable benefits to both institutions. For York University, the project will expand expertise in scalable, generalizable AI systems and interdisciplinary research in forestry, geography, and urban planning. For AIST, York University’s specialized skill set will strengthen the team’s capacity to tackle complex challenges in 3D data processing and advance the development of robust foundation models. This partnership will enhance research capabilities, foster innovation, and drive impactful advancements in AI and its applications.

Umaneo : Détection de défauts en 3D grâce à l’Intelligence Artificielle

Umaneo : Détection de défauts en 3D grâce à l’Intelligence Artificielle
Principales activités du partenaire et avantages escomptés du projet :
Umaneo est spécialisé dans le développement de solutions d’intelligence artificielle sur mesure. L’équipe d’experts et d’ingénieurs en IA d’Umaneo intervient à chaque étape du processus, du conseil stratégique à la mise en oeuvre de la solution. En utilisant la synthèse de données 3D, ce projet de stage vise à créer des ensembles de données complets englobant à la fois des modèles sans défauts et défectueux, facilitant ainsi une formation et une évaluation robustes des algorithmes de détection. La mise en oeuvre réussie de ce système devrait améliorer les processus de contrôle qualité, réduire les efforts d’inspection manuelle et améliorer l’efficacité globale de la fabrication.
Problématique :
Dans le secteur de la fabrication, des défauts non détectés peuvent entraîner des pannes de produits, une augmentation des coûts et une sécurité compromise. Les méthodes d’inspection traditionnelles sont souvent manuelles, longues et sujettes aux erreurs humaines. Le défi consiste à développer un système automatisé, précis et efficace qui exploite l’apprentissage automatique pour détecter les défauts en comparant les modèles 3D attendus avec les produits réels. Ce système devrait être capable d’identifier différents types et tailles de défauts, garantissant ainsi des normes de qualité élevées dans la production.

Tabular data extraction for financial disclosure analysis

Material Legal Technologies is a full-stack technology based legal advisory service that seeks to streamline the last mile in narrative public disclosure, in addition to developing various technological tools to alleviate friction in corporate governance.
This project aims to elevate our existing ML tooling to state-of-the-art in terms of recent research advancement, specifically helping with data intensive financial information extraction and parsing.
This project will help accelerate our growth and onboard new clients faster through improved data visibility and analytics, as well as serve our existing clients better with improved market comparative analysis. This will thus help drive revenue growth through client acquisition as well as improved and new product offerings.

A high throughput mass spectrometry-based screening test for early-stage colon cancer

Colorectal cancer (CRC) is the second most common cancer-related cause of death in Canada. This high mortality rate
is largely due to the lack of a cost-effective, patient-accepted and sensitive screening tool. Metabolomic Technologies
Inc. (MTI) has developed a nuclear magnetic resonance (NMR)-based urinary diagnostic test for detecting colonic
polyps (PolypDx™). Identification of adenomatous polyps can reduce colorectal cancer by 95%. PolypDx™ has a
sensitivity of 71% for detection of precancerous colonic polyps which is a significant improvement over the currently
available guaiac fecal-based tests with a sensitivity of 3-19%. With an NMR-based test, MTI’s market is limited to
facilities that have access to NMR. Developing a mass spectrometry (MS)-based test will expand the user market to
diagnostic laboratories. Through this application we propose to modify and adapt the NMR-based diagnostic tests to a
MS-based platform that will be high throughput, sensitive, and specific, and cost effective ($25-30/sample end-user
price).

Évolution des conditions hydrologiques de la plaine du Haouz (Maroc) en lien avec le climat, l’anthropisation et les séismes

La région du Haouz (Maroc), abrite l’un des aquifères les plus importants du pays et joue un rôle essentiel dans l’agriculture, le tourisme et l’industrie. Cependant, les sécheresses récurrentes, aggravées par les changements climatiques, entraînent une surexploitation des eaux souterraines. La pression croissante exercée par les activités humaines met en péril l’équilibre hydrique de l’aquifère, d’autant plus que des événements extrêmes, tels que le séisme survenu en septembre 2023, pourraient influencer la dynamique des eaux souterraines.
L’objectif de cette étude est d’évaluer l’impact des sécheresses et des séismes sur l’évolution des ressources en eau souterraine du Haouz, qui fait l’objet d’une exploitation croissante. Pour ce faire, une approche intégrée combinant des indices de sécheresse météorologique et hydrogéologique sera adoptée. L’Indice Standardisé de Précipitation (SPI) évaluera la variabilité des précipitations, tandis que les indices satellitaires, tels que l’Indice de Condition de Température (TCI) et l’Indice Différentiel Normalisé d’Eau (NDWI), permettront d’évaluer la disponibilité en eau.
Ces analyses, couplées aux Systèmes d’Information Géographique (SIG), à la télédétection et à la modélisation hydrodynamique de la nappe avant et après le séisme, permettront d’anticiper les changements futurs en vue de proposer des stratégies adaptées pour une gestion durable des ressources en eau.

Technical and business intern(s) working within cross-functional teams to commercialize AI-powered solutions (1)

AltaML builds artificial intelligence (AI)-enabled solutions to business problems. We work with organisations, bringing together their data and domain expertise with our AI expertise, to develop AI solutions that are deployed in their operations. We also commercialize AI-enabled products business via industry-specific ventures, yielding scalability from our investment in the first solution. AltaML’s AI Lab for Government, also known as GovLab, is a talent accelerator for public service professionals, post-secondary students and recent graduates. GovLab.ai’s mission is to set a global example of how to transform the public sector through applied AI, and is designed to encourage the growth of technical and business AI skill sets that are in high demand across Alberta and around the world. AltaML’s Venture Studio is an incubator program that works with founders and co-founders in the emerging tech industry to scale ideas, build venture products, and launch AI/ML startups across numerous industries.

Decentralized Electric Microgrid Optimization by Reinforcement learning

The goal of this project is to improve the management of electric community microgrids. In contrast to traditional power grids with large centralized power plants that provide energy in a top down fashion to consumers, the introduction of renewable energy such as solar has given rise to bottom-up electric microgrids of prosumers (i.e., consumers that also produce energy) where energy flows in a bottom up fashion from the edge of the grid. Furthermore, the intermittent nature of solar generation and its lack of synchronization with energy consumption creates important challenges for energy management. We will develop a decentralized agentic framework for grid management. More precisely, distributed reinforcement learning agents will dynamically manage energy production, storage and purchase/sale based on load measurements, weather information, demand patterns and spot prices. This decentralized agentic framework will help reduce peak loads, reduce costs, improve energy self-sufficiency, increase resilience to grid outages and scale to increasingly large communities of prosumers.