Projets novateurs réalisés

ML-optimization of tensor contraction operations (TCOs) in the PyBEST software package

The main goal of this project is to integrate advanced machine learning techniques into the open-source PyBEST quantum chemistry software package to substantially accelerate quantum chemical calculations through the automated selection of optimal computational strategies. The project specifically targets the AI-driven optimization of tensor contraction operations (TCOs), which constitute the primary computational bottleneck in many quantum chemistry methods. By enabling data-driven prediction of the most efficient contraction schemes and execution pathways across different problem sizes and hardware configurations, the proposed approach will significantly reduce time-to-solution, enhance scalability on modern GPU architectures, and improve computational resource efficiency, while preserving numerical accuracy and scientific reliability.

NouLife: Synthesis of Linked Conjugates

In 2023, Innovotech acquired a 60% interest in NouLife Sciences Inc., a company that developed the intellectual property for the linking of two antioxidant molecules commonly used in skin care products – alpha lipoic acid and acetyl-L-carnitine – into one molecule. The linking of the two molecules has been indicated by NouLife to increase the beneficial properties of the separate molecules by improving their penetration through the skin, thus promoting antioxidant activity and improving overall skin health as a result. Innovotech used the intellectual property to create three different conjugated molecules, and the next step is to confirm that the linking of the molecules provides more effective antioxidant activity in skincare applications, focusing initially on permeability, solubility, and antioxidant properties. This testing of the linked molecules will seek to confirm that the linkage itself significantly improves the function of the molecules with respect to skin treatment so as to encourage their use in nutraceutical and cosmetic applications. The proposed project is to work with collaborators at the University of Alberta to synthesize the additional quantities of the conjugated molecules needed for the next phase of testing.

L’origine leibnizienne du principe de la moindre action

Ce projet vise à rechercher dans la dynamique de Leibniz une nouvelle compréhension de la notion d’action de Maupertuis qui pourrait, croit-on, résister aux critiques des commentateurs qui ont suggéré que la quantité d’action est définie par Maupertuis de manière ad hoc et réhabiliter la preuve métaphysique de Maupertuis. J’avance l’hypothèse selon laquelle l’action de Maupertuis est une notion métaphysique qu’il est possible de préciser mathématiquement par l’usage de principes métaphysiques adéquats, de la même manière que Leibniz le fait dans Dynamica de potentia.

Scheduled Tandem Parking to Relieve Parking Demand at Commuter Organizations

This project develops and validates a tandem parking simulation and operating framework for large commuter organizations such as hospitals, universities, and municipal complexes facing persistent parking shortages. Conducted in partnership with pointA, a non-profit supporting sustainable commuter programs, the study explores how assigning two vehicles to a shared stall can safely and reliably increase effective capacity without new construction. Using shift-based data, stochastic modeling, and discrete-event simulation, the PhD intern will test pairing rules, buffer times, and participation rates to quantify utilization gains and blocking risks. The outputs—a simulation model, operational handbook, and planning guidelines—will provide evidence-based insights for future pilot implementation, helping organizations expand parking efficiency, reduce congestion, and improve access within existing footprints.

How Generative AI Changes Problem Solving in Software Development

Use of generative artificial intelligence (GenAI) has skyrocketed in software development. Although gains in workers’ productivity have received significant recent attention, little is known regarding the deeper long-term effects of these tools. Researchers from the University of Victoria, Canada and the University of Zurich, Switzerland are conducting an experimental study to evaluate the potential risks the technology poses to the critical thinking skills of software developers. The results of their research will help ensure GenAI is responsibly integrated into software engineering education curriculum at the participating institutions, and beyond benefiting mentors in the field, inform individual software developers.

Development and Characterization of Zr-Doped Multifunctional High-Entropy Alloy Coatings

This project aims to develop multifunctional high-entropy alloy (HEA) coatings that protect industrial components from multiple types of damage at the same time, including erosion, corrosion, and hydrogen attack. Using laser-directed energy deposition (L-DED), the research will study how adding zirconium to AlCoCr2FeMo0.5Ni –Zrx changes the coating’s microstructure and improves its performance for real industrial applications, such as aerospace, energy, marine, and hydrogen transport systems. The goal is to create long-lasting, reliable coatings that enhance the durability of critical components in harsh environments. Both participating institutions will benefit from shared expertise, access to advanced manufacturing and microscopy facilities, and opportunities for student training and collaborative research, strengthening their capacity to design and evaluate next-generation multifunctional protective coatings.

Studying collaborative interactions with physical props in extended reality

People in VR can see each other and see the same virtual objects, but the feeling of touch is usually weak or missing. It is not realistic to give everyone a full set of physical replicas for every virtual item they use. Instead, this project will explore how one or two simple physical objects can be reused in clever ways so they feel like many different virtual tools or parts. We will create a shared virtual space where two people, each in their own room with a VR headset, collaborate on tasks such as moving, aligning, or inspecting virtual objects. Both will hold only a small number of physical objects, but in VR, they will experience them as several different items that can be passed or shared. We will test various techniques to make this illusion feel natural and convincing, and then conduct a study to assess how well people can cooperate and when the illusion fails. The final goal is to offer clear, practical recommendations to help future VR tools make remote collaboration feel more physical, shared, and affordable.

Automated Structural Analysis and Change Detection for No-Code Automation Workflows

This project focuses on developing and evaluating automated methods for analyzing complex no-code automation workflows used in enterprise environments. Platforms such as Workfront Fusion and Make.com allow organizations to build powerful automations quickly, but as these workflows grow in size and importance they often become difficult to understand, maintain, and safely modify. Changes are frequently made without clear visibility into downstream impacts, which increases operational risk and makes long-term governance challenging. The project will investigate techniques for parsing automation configurations, identifying structural relationships within workflows, and detecting meaningful differences between versions over time. In addition, the project will explore ways to generate clear, human-readable summaries and visual explanations that help technical and non-technical stakeholders understand how automations function.

The expected benefit to the partner organization is improved clarity and control over their automation systems. By making the structure and behavior of complex workflows easier to understand, the organization can reduce the risk of errors when changes are introduced, support more informed technical decision-making, and improve collaboration between developers, administrators, and business stakeholders. These capabilities will also help the organization scale its use of automation more responsibly by supporting better documentation, auditing, and long-term maintainability. Overall, the project aims to transform opaque automation configurations into understandable, manageable systems that can evolve safely as organizational needs change.

Détection, prévision et interprétabilité des sécheresses extrêmes à l’aide de l’intelligence artificielle

Les sécheresses extrêmes constituent un enjeu majeur dans un contexte de réchauffement climatique, affectant l’agriculture, les ressources en eau, la production énergétique et les écosystèmes. Leur caractère multi-échelle, dépendant des interactions entre précipitations, évapotranspiration, humidité du sol et stockage souterrain, rend leur détection et leur prévision complexes. Les indices traditionnels (SPI, SPEI, SSI) apportent une évaluation utile, mais ils peinent à représenter la dynamique complète du système hydrologique, en particulier dans les régions à disponibilité limitée de données.

Les avancées en télédétection améliorent la surveillance, mais nécessitent des méthodes capables d’intégrer des données hétérogènes. L’intelligence artificielle (IA) offre ici un potentiel important. Des modèles tels que Random Forest ou LSTM ont montré de bonnes performances de prévision, mais leur manque d’interprétabilité limite leur utilisation opérationnelle.

L’intégration d’IA explicable, notamment via des approches comme SHAP, permet de lier les prédictions aux processus physiques. Couplée aux projections climatiques CMIP6 (scénarios SSP), elle offre la possibilité d’anticiper l’évolution future des sécheresses, en identifiant les facteurs critiques, tels que l’augmentation de l’évapotranspiration. L’objectif est de développer un cadre intégré pour améliorer la gestion et la résilience des territoires face aux sécheresses.

M.I. Understanding Social Platform Strategy Optimization

M.I. Understanding is a community-centered educational organization that supports families, caregivers, educators, and children as they navigate important social and developmental challenges. Through online learning modules, original puppeted educational videos, and curated guidance to external supports, the organization builds understanding around childhood mental health, PRIDE education, and an upcoming expansion into Fetal Alcohol Spectrum Disorder (FASD). Delivered in partnership with schools, libraries, and community agencies, its programs foster early, supportive conversations and provide families across Canada with accessible tools for emotional resilience, skill building, and navigating complex developmental needs. Because this mission relies on cultivating an informed and connected community, establishing a strong, intentional, and evidence-based online presence has become essential for the organization’s long-term impact and sustainability.

This project will support M.I. Understanding by developing a comprehensive and optimized social media strategy designed to expand reach, strengthen engagement, and make its educational content more accessible to diverse audiences including families, educators, and potential partners. Over four months, the project will accomplish four core objectives: 1) assessing current platform performance; 2) defining overarching social media goals aligned with the organization’s mission; 3) creating platform-specific strategies tailored to each audience; and 4) launching, testing, and refining a prototype content plan. Guided by the Toronto Translational Thinking Framework (TTF), the project will follow an evidence-based, iterative approach that moves from observation and data collection to analysis, contextualization, solution design, prototyping, and refinement.

The project will enhance the M.I. Understanding’s capacity to deliver mental health education, foster community support, and promote early skill development by translating its existing resources into compelling, accessible social media content. A stronger digital strategy will help the organization reach new families, engage existing users more meaningfully, and broaden its impact across Canadian communities. Ultimately, this work will provide M.I. Understanding with a framework that supports its mission to foster understanding.

Advanced characterization of adhesively bonded CFRP-to-concrete systems

In the context of strengthening existing structures, the use of fibre-reinforced polymer materials has been continuously increasing as an alternative to traditional materials, due to their superior durability (absence of corrosion), lightweight, low maintenance cost, and rapid installation. Adhesive bonding techniques for strengthening existing structures, such as near-surface mounted (NSM) and externally bonded reinforcement (EBR) are usually preferable. Despite the increasing knowledge, the long-term performance of these strengthening techniques under hygrothermal ageing conditions is still not clearly understood, limiting their use. This work aims to give new insights for reliable predictions of the long-term performance of these strengthening techniques, supported by advanced and innovative numerical modelling, properly calibrated with a refined experimental program that will be developed in this work. These outcomes will provide new knowledge on the most influencing parameters, serving as a basis for the development of simplified design recommendations.

L2M – NeviSight

NeviSight is an AI-powered platform that transforms retinal imaging into faster, more accurate, and more accessible eye disease detection. Designed as a decision-support tool, it aims to integrate seamlessly with existing fundus cameras and clinical workflows, providing clinicians with real-time lesion segmentation and risk insights. Our completed prototype combines a proprietary hybrid AI model with a user-friendly UI/UX, validated in partnership with ocular oncologists from the Alberta Ocular Brachytherapy Program. By advancing toward Health Canada and FDA approval, securing strategic partnerships, and refining deployment in clinical settings, NeviSight is positioned to reduce preventable blindness and fatal cancer through digital health innovation.