Innovative Projects Realized

Tensor based machine learning with human computer interaction applications

In this research, first we will using tensor network techniques speedup the processing in neural network since computational cost is a major bottleneck in neural network based deep learning. Note that the weight matrix in each layer of neural network is with huge size, because of millions of parameters. This may cause large time complexity to calculate all the parameters. To overcome the large time complexity, we could use tensor decomposition to calculate the low-rank weight matrix, to reduce a large amount of parameters, therefore to reduce the time complexity. The tensor networks include Hierarchical Tucker (HT), Tensor Train (TT), CP, and Tucker decomposition etc. Second, we propose to develop tensor based algorithms to solve spatiotemporal corrupted problem, complex background problem in human action recognition (HAR) and denoising problems in medical images. The human action videos are represented as a high-dimensional tensor, and we can solve the spatiotemporal corrupted problem by tensor completion, and recurrent neural network to deal with the missing frames problem; also we could denoise the high-dimensional medical images by tensor completion.

Assessment of indoor environmental quality in the school building

Indoor environmental quality in schools is one of the major issues that should be considered. Indoor air quality (freshness and cleanness of air inside the building) and thermal comfort (the comfortable condition when no one is feeling either too hot or too cold) in schools can have significant impact on children’s health, learning and productivity. The purpose of this research is to evaluate indoor environmental quality (indoor air quality, thermal comfort, lighting, and acoustics) in First Nation School. To achieve this goal we conduct long-term measurements of indoor air quality and thermal comfort parameters (CO2, temperature, and relative humidity) and ask the teachers to participate in voluntary and anonymous questionnaire survey about their perception of parameters of indoor environmental quality. The data will be shared with the partner organization to develop different retrofit packages to improve indoor environment quality and energy efficiency of the school building depending on potential to achieve goals such as energy savings, CO2 reduction, cost savings, improvement in thermal comfort by using Building Information Modeling optimization tool.

Predicting 30-Day Risk of Hospitalization, Emergency Department Visit and Death Among Albertans Receiving Opioid Prescriptions Using Machine Learning Algorithms: A Prognostic Study

Prescribers rely on guidelines for appropriate use and prescribing of opioids. Currently, there are no risk prediction calculators available to quantify the risk of prescribing an opioid that uses a patient’s history and drug utilization pattern. Machine learning algorithms offer an opportunity to perhaps develop such a risk calculator as a result of the large amount of heath data being collected in Alberta. OKAKI, a clinical analytics company, may benefit from this project as it may be able to offer a risk calculator to prescribers as a clinical decision aid tool. This may ultimately enhance patient safety.

Ultra-thin graphene oxide membranes for efficient humidity harvesting (Part II)

The main goal of the proposed work is to develop an ultra-thin and selective GO membrane capable of separating water vapor or steam from air. For this purpose, a suitable membrane supports required to hold the GO sheets. Therefore, the GO sheets will be deposited on various membrane supports and their performance in terms of selectivity, permeability, and mechanical strength will be evaluated. Then, the effect of GO layer number on the selectivity and permeation rate will be investigated. Once the best performing membranes have been determined, the impact of feed humidity and temperature on the permeation rate and selectivity of water vapor transport through the membranes will be evaluated. Evercloak will supervise the intern and support out of lab research activities on their premise. Evercloak will benefit as the results of this project will inform key decisions within their technology development roadmap in addition to talent acquisition/training.

Developing and evaluating multi-functional rain gardens

Rain gardens are shallow planted depressions in the ground that infiltrate, and evaporate urban rainwater. They are a kind of green infrastructure that provide an alternative to putting polluted urban runoff into the storm sewer system. The researchers will build and test new designs and installation methods for rain gardens. The new designs will incorporate habitat for bees and other pollinators and will use a mix of planted and seeded plants. The public will be shown pictures of the research rain gardens in an urban setting and asked to give their opinions as to which ones they prefer. The costs of installation and maintenance of each of the rain gardens will be recorded and compared.

The Hidden Subgroup Problem in Post-Quantum Cryptography

Quantum computers are new technologies that harness the power of quantum mechanics to store and manipulate information. Recent years have seen accelerated progress on their development, and while their potential benefits are significant, we must now prepare the cryptographic infrastructure underlying cybersecurity to ensure the continued safety of our data in the face on an adversary with access to a quantum computer.
Sectigo, a world leader in cybersecurity, together with researchers at the University of Ottawa, are teaming up to ensure that the latest academic research is integrated into the cutting-edge cryptographic protocols they deploy. The key focus of their work is ability of a core mathematical problem – called the Hidden Subgroup Problem – to withstand attack in the face of a quantum adversary, and what this means to our post-quantum cryptographic contenders.

Litho-minéralogie automatisée de carottes de forage et de poudres de minerais pour des applications géométallurgique et géoenvironnementale

Le projet présenté ici vise le développement d’une technique de caractérisation lithologique automatisée adaptée pour les carottes de forage en exploration minière. La technique sera basée sur divers capteurs (optiques, chimiques et autres) ainsi que la microscopie optique. Les données obtenues (big-data) serviront à l’optimisation des procédés de traitement des minerais ainsi que la bonne gestion environnementale des rejets issus des exploitations minières.

Safe Magnetic Resonance Imaging of Interventional Devices

Magnetic resonance imaging (MRI) is a non-ionizing modality; therefore, there is no

known damage on human tissue as opposed to ionizing imaging modalities, such as

X-ray. Interventional procedures take long time and in order to prevent ionization,

interventional MRI is very promising. On the other hand, there are safety concerns

because of the possible temperature rise on current carrying conductive wires, such

as guidewires and biopsy needles, during MR scans when it is inserted inside the

human body. In this project, we are aiming to eliminate the dangerous temperature

rise by using different methods and improve safety of the devices used in

interventional MRI.

Habitat security mapping for improving human-wildlife coexistence in the Bow Valley

Tourism, nature-based recreation and residential development within the Bow Valley (BV) of Alberta’s Rocky Mountains are all valued and continually increasing. The BV is also important for large mammals as it provides habitat for iconic species like grizzly bears to live in and move through. The Yellowstone to Yukon Conservation Initiative (Y2Y) is dedicated to balancing the needs of people and wildlife, and collaborated with stakeholders to recommend how to improve human-wildlife coexistence in the BV. Targeting one of those recommendations, this research project will help wildlife managers and stakeholders by developing a method for identifying areas of high-quality, undisturbed grizzly habitat. Land managers can use this information to manage human use in these areas accordingly. By establishing methods that help balance the needs of people with wildlife, Y2Y and the people of the BV continue to exemplify efforts to maintain wide-ranging populations both inside and outside of protected areas.

Violence and Smuggling Markets: State-smuggler interactions on the Brazilian border

Different states’ laws and law enforcement capacities meet in border areas. The legal and regulatory differences, the distinct police and administrative capabilities involved, as well as the degree of commitment of state authorities, create rent potentials that are central to border dynamics and the frequent instability that prevails in border regions. This has been traced to the smuggling of legal and illegal goods and to the peculiarities of military and police enforcement of the movement of people and goods. On average, border violence in Brazil from 2000 to the present has been higher than in the rest of the country and concentrated in municipalities with smuggling markets. However, in these markets there have been large changes in the levels of violence over time and across municipalities without obvious reasons for such variation. Those levels of violence, however, have been highly volatile over time. The proposed research thus seeks to understand why, and under what conditions smuggling markets become violent, focusing on the interaction of state authorities with non-state actors.

Développement d’un système d’aide à la détection et à la localisation des défauts présents sur les photographies aériennes par intelligence artificielle

Le projet proposé vise à développer un système d’aide à la détection et à la localisation des défauts présents sur les photographiques aériennes par intelligence artificielle, et plus particulièrement par l’utilisation des réseaux de neurones convolutifs. Ces derniers sont reconnus pour leur performance en vision par ordinateur, notamment dans la détection automatisée d’éléments fournis à travers des exemples d’entrainement. Le système d’aide à la détection proposé sera composé de trois architectures d’apprentissage profond complémentaires pour répondre aux problématiques rencontrées par la compagnie dans le processus d’analyse des photographies aériennes : (1) la détection des nuages et/ou des ombres ; (2) la détection du flou ; et (3) la détection de huit défauts majeurs connus lors de la construction des orthophotos et des orthomosaïques. Les algorithmes développés seront intégrés dans un système opérationnel qui permettra d’automatiser une bonne partie de la phase d’analyse et ainsi de réduire le temps et les coûts de cette opération pour la compagnie.

Assessing Equipment and Technologies for use in the Development of Dykelands and Upland in Atlantic Canada for Sustainable Agricultural Production

The research in this Master’s project is focused on documenting and analyzing the contemporary costs associated with maintaining and/or developing agricultural dykeland and the costs associated with the development of upland. It will also recognize the evolution of land development equipment and identify modern equipment and technologies appropriate to maximize agricultural potential on both dykeland and upland. Primary outcomes from this research will involve determining the financial cost of improving under-utilized dykeland and a return on investment using contemporary technology and innovative techniques versus traditional methods. This will involve researching, and potentially testing, cutting-edge technologies used in agriculture and dyke maintenance including, unmanned aerial vehicles (UAVs), structure from motion (SfM) software, remote sensing, topographical surveying, land development machinery and precision agriculture. These technologies will be subjected to a cost-benefit analysis to determine their efficacy in modern dykeland management.