Innovative Projects Realized

Development of prediction models in smart buildings using data driven approaches through Artificial Intelligence

BrainBox AI is working on the mandate to provide better solutions for optimized control of HVAC systems modeled through data using AI. The Intern will continue our research at BrainBox AI and to continue to develop data driven approaches for an optimized operation of HVAC systems in design and implementation of scalable AI frameworks for AI needs in the ongoing projects at BrainBox AI. The Intern will apply AI (including ML/DL/RL), data mining and statistical approaches for the creation of scalable predictive models. Evaluate the performance of predictive models and integrate in existing AI frameworks identify, analyze, and interpret trends or patterns in complex data sets. The solution to achieve this objective will be created by applying AI algorithms.

Analyse des réponses gravimétriques en forage du système Gravilog

La gravimétrie est une méthode de géophysique appliquée qui utilise la loi d’attraction

Universelle pour sonder le sol. Cette méthode est utilisée pour calculer les changements de

densité des roches. Puisque les gisements miniers de la région de l’Abitibi sont beaucoup

plus denses que la roche l’entourant, il est possible de les détecter à l’aide de cette méthode.

Récemment, un nouvel instrument de mesure, le Gravilog, a été développé par la compagnie

Scintrex. Le Gravilog est un gravimètre qui mesure la variation d’attraction dans un forage. À

l’aide de cet instrument, il est possible d’observer des gisements profonds et de mieux les

positionnés. En plus, la méthode gravimétrique peut être utilisée en complémentarité avec

d’autres méthodes géophysiques en forage déjà utilisées par Abitibi Géophysique pour mieux

positionner et dimensionner un gisement. Cependant, l’utilisation de cet instrument est

coûteuse, alors il faut bien optimiser son utilisation. Pour ce faire, trois aspects seront

analysés.

Graph-based learning and inference: models and algorithms

Learning from relational data is crucial for modeling the processes found in many application domains ranging from computational biology to social networks. In this project, we propose to work on developing modeling techniques that combine the advantages of the approaches found in two fields of study: Machine Learning (through graph neural networks) and Statistical Learning (through statistical relational learning methods). By combining the advantages of both approaches, we aim to obtain better prediction results for an array of problems such as classification and link prediction in relational data.

Unsupervised Learning of 3D Scenes from Images using a View-based Representation

We’d like to address the issue of 3D reconstruction from 2D images. This means developing a machine learning algorithm that can take a regular photo as an input and generate a full 3-dimensional reconstruction of the contents of the photo. Such technology can be used creatively or to help the coming generation of robots better understand their surroundings.

Surface property optimization of an insulation coating for glass

The handling and storage of perishables like vaccines are critical to retain medical potency and effectiveness for patient administration. The delivery of vaccines can succumb to several pitfalls such as physical damage, losses in transit and expiration. One of the primary causes of vaccine wastage is a result of heat and freezing of the biological products. These sensitive perishables quickly spoil if not maintained in the correct temperature range of a refrigerator that is kept to +2°C to +8°C. [1] As a result of vaccines being exposed to temperatures outside of this 6-degree range can reduce the protection that vaccines provide against such diseases like diphtheria, pertussis and tetanus. Clear visual differences can be observed from within vials that show large precipitates of crystalline structures of the affected vaccine when exposed to freezing temperatures. [2] These unfavorable conditions irreversibly ruin the vaccine for immunization. Ensuring proper vaccine temperature conditions in-transit have mainly been done through cold-chain delivery and the use/management of expensive purpose-built refrigerators.

Les didacticiels de films interactifs : entre intégration diégétique et instructions procédurales

Se situant dans la lignée des études du jeu vidéo en sciences humaines, mon projet de recherche prend pour objet les didacticiels, c’est-à-dire les niveaux ou instants de jeu “apprenant” à l’utilisateur comment interagir avec le logiciel (“appuyez sur la touche X pour sauter”). Il se concentre sur le genre du cinéma interactif, afin de replacer des œuvres contemporaines – comme “Black Mirror : Bandersnatch” (David Slade, 2018) sur Netflix – dans la filiation des films interactifs nommés FMV (Full Motion Video). Ces productions hybrides, mêlant interactivité et images en prise de vues réelles, sont apparues au sein de l’industrie vidéoludique il y a bientôt trente ans, avec des œuvres brouillant chacune à sa façon les frontières entre cinéma et jeu vidéo. Ces “films-jeux” et leurs manuels, boîtiers et autres notices d’emploi sont conservés au LUDOV (Laboratoire Universitaire de Documentation et d’Observation Vidéoludiques) de l’Université de Montréal. Ma recherche permettra de redécouvrir ce corpus et d’analyser, à partir de leur(s) didacticiel(s), les stratégies employées par les développeurs pour s’adresser au joueur/spectateur et nouer des “contrats de lecture” ludiques, dont on étudiera toutes les modalités et paradoxes énonciatifs.

The Impact of Emerging Technologies on Defense Policy

How technology relates to Canadian national defence policy, strategy, and planning is not well understood. To date, much of the research on emerging technology in the military focuses on concerns over ethics, privacy, human rights, and governance, largely overshadowing equally important research on the way technology intersects with military strategy and tactics, defence recruitment and training, and alliance politics and national policy. Our project is inherently forward looking: its overarching objective is to think creatively about emerging technology and to link developments therein to Canadian defence.

Lay Sisters and Brothers in Medieval German Convents

Medieval nuns were supported by a wide range of people, including lay sisters and brothers who performed much of the physical and manual work of the convent. Yet scholarship has long ignored these members of the monastic community. My research attempts to address this gap by focusing on lay sisters and brothers in fourteenth- and fifteenth-century German convents. I would like to request funding to spend three months in southern Germany next summer, from the beginning of May 2020 through to the end of July 2020. During this time, I will search through archival documents from medieval convents, including deeds, visitation records, account books, necrologies, and convent statutes, to find references to lay sisters and brothers in various archives in southern Germany. I am a third-year PhD student at the Centre for Medieval Studies in the University of Toronto, and this ground-breaking research will help to nuance our understanding of women’s daily lives, religiosity, and the intersections of gender and class in medieval Europe––subjects that have historically been neglected by historians. I believe that undertaking this research next summer will provide me with the remaining research necessary for my dissertation and greatly contribute to the study of female monasticism.

Health Equity and Barriers to LDKT in South Asian Communities

As of 2012, about three million Canadians suffer from chronic kidney disease. Patients with kidney failure need dialysis or kidney transplantation to survive. LDKT (LDKT) is the best treatment option for many patients with kidney failure, however, it is not used as much as it could be. Patients who belong to ethnic minority groups are less likely to receive LDKT compared to Caucasians. Currently, there is limited research to help us better understand what prevents access to LDKT in patients from the South Asian community. The purpose of this research is to engage the South Asian community to understand social, cultural, and religious barriers to LDKT while also raising awareness of kidney disease and LDKT. This research will also engage different stakeholders to help develop a Health Equity Strategy for the South Asian communities in Ontario.

High temperature wear of sand cast and die cast Al and Mg alloys for continuously variable transmission (CVT) and other applications

The project will determine sliding wear behaviour of lightweight Al and Mg alloys for the continuously variable transmission (CVT) bore application and other engine applications. High temperature lubricated wear tests that simulate the actual operating conditions of CVT bore will be carried out. In addition, tests will be conducted at room temperature and under the dry conditions to benchmark wear resistances of different cast alloys and determine the wear mechanisms. Tests will be conducted using a block-on-ring tribometer at the University of Windsor’s Tribology of Materials Research Centre. The project will simultaneously determine the roles of hardness, roughness, porosity as well as tribolayer formation on the sliding wear behaviour of the alloys by conducting worn surface and subsurface characterizations using analytical microscopy and spectroscopy. The results will be presented in the form of wear mechanism maps, which will be used to develop new wear resistant lightweight alloys by GM.

Development of a macroscopic model to simulate water beads on hydrophobic andsuperhydrophobic coatings to be used in-flight icing software such as CANICE-2D

The intern has developed a 3D microscopic model to simulate water bead behavior on

hydrophobic and superhydrophobic coatings. The microscopic model developed at Concordia

University which simulates one bead cannot be directly integrated into an aircraft icing panel

code such as CANICE-2D. CANICE-2D approved by Transport Canada has been widely

used at Bombardier Aerospace to simulate aircraft in-flight icing. This is a panel code solving

the flow, the droplet trajectories, ice thermodynamic and the geometry. CANICE-2D can run

cases in few minutes which reduces the time to find the critical icing shape. CANICE-2D

calculates the ice shape by solving the heat and mass transfer equations and the water film

at the surface when glaze ice is formed.

Adding value from coffee waste: low oxidized sterols in silverskin

Coffee production with imported beans has doubled in Canada since 2003. From this > 125,000 t of waste is generated most of which ends up in landfill. A significant component of the waste is as the pulp (cascara) and silverskin. In 2019, the Happy Goat Company roasted 50 t of green coffee beans yielding 300kg silverskin which will greatly increase in 2021. Silverskin teas have 3-6 times more caffeine than coffee beans and are rich in total phenolics, antioxidant capacity and phytosterols. There is good evidence of their safety. However, the coffee roasting process creates oxidized phytosterols in the silverskin which are somewhat toxic. Phytosterol Oxidation Products (POPs) have been reported in silverskin in several studies. We have a unique opportunity to examine POPs in silverskin considering a number of variables. These will include process, roast degree and the roast profile. When the final best conditions are chosen, measurement of caffeine, total antioxidant capacity and total flavonoids with be conducted in the final product and in the silverskin-derived beverages. The goal is to reduce waste from the roastery and add some value for the farmers in Ethiopia and Guatemala who produce coffee for the company.