Innovative Projects Realized

Multimodal Representation Learning for Pregnancy Care

Machine learning for healthcare promises to improve our diagnostic ability, to plot optimal treatment plans, and to smooth clinical operations and lower costs. But, central to these promises is a strong, generalizable numerical representation of a patient’s prior medical history that is reflective of clinical similarity and amenable to modeling. Learning such a representation is difficult as health data, including EHR data, is notoriously messy, sparse, and multimodal (e.g., comprised of many different types of information streams). In this work, I will build a machine learning system capable of learning a suitable representation of a patient, principally by leveraging massively-multitask modeling and structured methods such as graph networks. This representation will be evaluated through its applicability in pregnancy care, specifically through its ability to perform generalizable risk profiling across key outcome measures, such as low birth-weight and pre-term birth, and through its ability to perform patient subtyping.

Integrated computational intelligence for forecasting import and export volumes of commodities

The global seaborne trade is growing rapidly and being supported by shipping, one of the major transportation tools around the world. With the satellite and terrestrial AIS (Automatic Identification System) data, it is possible to track the trajectories of vessels carrying commodities. The capability to accurately forecast the import and export volumes and types of commodities will potentially enable the maximization of business trading profits. This research aims to develop forecasting algorithms to predict the trading in the business with the vessel tracking and cargo inspection data. Integrated with industry data analytic platform, the outcome from this research will enhance Canadian business’ competency in the global market.

Evaluating the influence of water fill level on the retort come-up time, temperature andheat distribution in water immersion/spray retorts

Thermal processing is a major method of food processing and involves heating of foods packaged in

hermetically sealed containers to destroy pathogenic and spoilage microorganisms in the product.

While safety and shelf-stability of thermally processed foods were the primary concerns of consumers

in the past, today the consumer demand higher quality processed products. Rotary autoclaves are

employed for such purposes since they provide more rapid heat transfer and reduce the cook time.

Shorter retort come up times to operating conditions offer better quality retentions since they will

reduce the overall heating time. The purpose of this project is to evaluate the influence of retort water

fill level on the come-up time and temperature/heat distribution during water immersion rotary

autoclaving. The study has the potential to demonstrate alternate processing procedures which may be

beneficial in tenns of reducing the come-up time or improving the temperature and heat distribution

leading to better quality products.

Optimization of Angiotensin II Receptor type 1 Blockers (ARBs) in chronic obstructive pulmonary disease (COPD)

Chronic Obstructive Pulmonary Diseases (COPD) is a lung disease that cause a lot of suffering to the Canadian population. To accelerate the drug discovery process, an old blood pressure lowering medication was tested to block the progression of COPD. A patient study showed that the old medication did provide some protective effect to the lung airways of COPD patients. However, we have found that this old medication does not slow down COPD by lowering blood pressure, but rather by acting on a new, unknown target. As expected, we found that this old medication is an ‘average’ mediator of lung protection, and there are other medications in the same class of drugs that may provide better lung protection. We will assess these other medications and will also make some modifications to the structure of the old medication to improve its ability to protect the lung airways of COPD patients. Our innovative proposal to speed up the drug discovery process will benefit COPD patients and the management of their disease.

Airway-On-A-Chip: Development and In Vitro Validation of A Microfluidic Cell Culture Model for Chronic Obstructive Pulmonary Disease (COPD)

Chronic obstructive pulmonary disease (COPD) is an inflammatory disorder of the lung, and one that affects 2.6 million Canadians and 380 million people worldwide. Although the disease affects a large population worldwide the therapies used for treatment remain imprecise. With the lack of disease modifying therapies there is a pressing need to discover novel targets to promote new therapeutic discoveries and ultimately improve the care and health outcomes of patients with COPD. To date, the discovery of novel therapeutics has been greatly hindered by outdated cell models and costly animal models. Our group proposes that through collaboration between engineers, pharmacologists, biologists and physicians we can leverage Organs-on-chips (OOC) technology to discover and validate novel COPD therapeutic targets. Organs-on-chips are miniature devices mimicking aspects of the in vivo conditions of human organs. What we propose to do in this project is to build upon a previously developed airway-on-a-chip technology to better mimic the human airways.

AI-Based Automation of the Candidate Recruitment and Management System for TAS

TAS (Techno Aero Services Inc.) is a Canadian recruitment agency specialized in aeronautics, engineering, and technology. At TAS, candidates currently submit resumes to a database through a web interface. These resumes are then manually processed by recruiters before a suitable candidate is matched to a position. Through this laborious manual process, great prospective candidates often get lost in the piles. This project aims to leverage text mining and machine learning techniques to automatically collect information about prospective candidates from their submitted resume and profiles on social and professional networks like LinkedIn, Twitter, and Facebook to provide recruiters with an overall picture of a candidate’s strengths and weaknesses. Both the technical and communication skills of the candidates will be analyzed and summarized. The goal is to provide recruiters with the information they need to judge a candidate on both his mastery of key skills and his ability to fit into the company culture. In addition, we aim to leverage the existing data of the candidates and their employment records to build Business Intelligence (BI) component for the proposed system to predict business trends, support business decisions and to reach and recommend prospective candidates for available jobs in TAS’s job pool.

Classification de produits pour l’optimisation des stratégies d’approvisionnement chez Stelia

Stelia Amérique du Nord (Stelia) est une entreprise du secteur aérospatial qui offre différents services, produits et composants de haute technologie aux grands donneurs d’ordres comme Airbus et Bombardier. Stelia opère deux importantes usines en Amérique du Nord et achète globalement environ 14.000 pièces et composants spécifiques et standards. L’objectif de ce projet est d’aider Stelia à optimiser la gestion de ses approvisionnements en définissant des classes de produits qui soient adaptées à des stratégies d’approvisionnement spécifiques et efficaces. Pour cela, nous proposons d’utiliser la simulation à événements discrets et l’analyse de données par arbres de classification pour définir des règles de classification de produits à partir des résultats de simulation de stratégies d’approvisionnement de ces produits (ex., coût d’approvisionnement, retard moyen, niveau de service). En particulier, Stelia souhaite évaluer la performance potentielle de stratégies d’approvisionnement avec fournisseurs multiples de manière à gérer les aléas de sa consommation, ainsi qu’une stratégie avec un entrepôt de consolidation au Canada.

Modélisation du mécanisme des décharges partielles d’encoche en fonctionde la dégradation des surfaces

La fiabilite des alternateurs hydrauliques est un sujet d’importance pour les utilitaires afin

d’assurer la fiabilite de I’approvisionnement en energie electrique. La maintenance de ces

equipements est donc un outil important dans la gestion et I’exploitation d’un parc

d’alternateurs. Apres un certain nombre d’annees de fonctionnement, I’usure de ces

equipements peut mener a I’apparition de decharges partie lies entre I’isolation de masse des

bobines et Ie circuit magnetique. Le sujet de ce stage est I’etude des mecanismes relies a

I’apparition de ces decharges partielles ainsi que I’etude de I’evolution de la degradation

occasionnee par ces decharges partielles.

Développement de modèles prédictifs pour la détection de défaillance sur les pièces prioritaires de turbines éoliennes.

La société Enercon, produit des éoliennes et par la suite opère les centrales qui ont été installés, pour garantir le
bon fonctionnement de celles-ci. L’un des enjeux à surmonter pour la maintenance d’une centrale éolienne est la
prédiction des défaillances, afin d’intervenir le plus rapidement possible et de limiter le temps d’arrêt des turbines.
La compagnie souhaite donc instaurer des indicateurs de mesure basés sur de l’intelligence artificielle, qui
permettront un remplacement plus efficace du matériel mécanique, afin de diminuer les coûts de maintenance et
d’augmenter la production sur le cycle de vie d’une éolienne. L’objectif de ce projet de collaboration est la mise
en place d’algorithmes se basant sur l’historique des maintenances et les mesures effectuées depuis la
construction de la centrale afin de prédire les futurs comportements des pièces prioritaires d’une éolienne comme
les pales, la génératrice, le transformateur électrique.

Application of advance signal and image processing to develop objective diagnostic and monitoring technologies as well as predicting the response to treatment for Alzheimer’s disease

An ongoing study in Winnipeg is investigating the possibility of using repetitive transcranial magnetic stimulation to help treat the cognitive and memory declines in Alzheimer’s disease. The objective of this project is to analyze in detail the current results of this study. We will look for connections between MRI scans of the participants and their response to the treatment. We will also look at anxiety and resting motor threshold (a measure of how sensitive they are to our treatment device), to see how they change the effectiveness of treatment. We will also look in detail at all of the additional measurements we performed as part of the study to see if any unexpected results can be found. Finally, we will look for changes in brain activity as measured using electrical signals from the balance organs in the ear.

Caractérisation génétique de souches sauvages d’algues (Saccharina latissima), issues de régions géographiques différentes au Québec et tests en algoculture de souches d’intérêt sélectionnées.

Au Québec, les fermes marines ayant un intérêt pour l’algoculture sont situées aux Îles-de-la-Madeleine, en Basse-Côte-Nord et en Gaspésie, sur des sites dont les caractéristiques sont très différentes. À cet effet, une entreprise maricole de la Gaspésie, Ferme Marine du Québec, produit commercialement des plantules sur cordes et vend des collecteurs d’algues à des aquaculteurs désirant faire la culture d’algues. Pour le moment, l’ensemble des entreprises algocoles au Québec utilise des plantules provenant de géniteurs de Bonaventure en Gaspésie. L’introduction d’espèces cultivées est l’une des menaces les plus importantes relatives à l’aquaculture, responsable d’effets négatifs sur les écosystèmes. Malgré son importance écologique et économique, la connaissance de la génétique des populations de l’espèce S. latissima au Québec est inexistante. Le stagiaire vérifiera si les populations d’une même espèce sont génétiquement distinctes.

Photoacoustic remote sensing microscopy for pre-clinical and clinical applications

There is a desperate need for a fast, non-contact, non-invasive, safe and accurate technique that can measure oxygen saturation, oxygen metabolic activity and multilayered histology-like information. Oxygen saturation and oxygen metabolic activity play a vital role in understanding several diseases including early tumour and vision loss diagnosis and treatment. Additionally, when the oncologic care team must to remove a tumour, it is essential that no malignant tissue left behind. The ability to predict tumour aggressiveness, margin and metastatic potential could significantly affect clinical practice in oncology. It is also shown that abnormal retinal oxygen saturation and metabolic activity are the leading causes of vision loss (e.g. age-related macular degeneration, diabetic retinopathy and glaucoma). The ability to precisely detect aberrant retinal oxygen metabolic activity in a non-contact setting is essential for improving investigations and diagnoses of ocular diseases. In addition, for various applications is surgical oncology and optometry contact is impractical, or the working space and footprint is an issue (e.g. endoscopy and surgery). To address these needs, Mitacs funds are requested to recruit HQPs which is critical to the advance of Histological and Functional Photoacoustic Remote Sensing (HF-PARS) microscopy.