Projets novateurs réalisés

Representing and understanding 3D Objects

Understanding the 3D world around us is a key goal in computer science. If we are endowed with 3D understanding, we can solve problems in several areas in computer science such as robotics, fast 3D modelling and physical reasoning. However, it is not clear how 3D objects can be represented because of the high dimensionality as well as different sources of getting 3D information. Furthermore, in a commercial setting, there are many stages of the graphics pipeline which are very tedious and automating certain parts can significantly simplify the process. The goal of the research is to investigate different methods to represent the shape and structure of 3D objects using deep learning techniques. This knowledge can then be used to create algorithms which can easily do 3D object classification, segmentation and generation. This research will hopefully lead us to have a better understanding of 3D representation which has countless research and commercial applications.

Création de valeur dans les projets d’efficacité énergétique de bâtiments grâce aux données d’opération

Au fil du temps, les bâtiments génèrent de grandes quantités de données d’opération. À l’heure actuelle, celles-ci sont peu ou pas exploitées. Pourtant, les différents acteurs de l’industrie sont d’avis que ces mégadonnées pourraient cacher des informations utiles pour améliorer la conception et l’opération des différents systèmes du bâtiment, mais les meilleures façons de le faire ne sont pas encore établies. Le présent projet vise à étudier les bénéfices potentiels liés à l’exploitation de mégadonnées provenant de projets en efficacité énergétique, comme ceux réalisés par le partenaire du projet Ecosystem. L’utilisation de l’information latente contenue dans les données pourrait permettre de réaliser plus d’économies d’énergie et créer de la valeur ajoutée aux projets d’efficacité énergétique.

Deloitte IP Factory: dTRAX Contract Management

Manual contract analysis and management is a laborious task. dTrax is Deloitte’s managed solution to this problem—it uses machine learning to automate the arduous contract management process and help users gain further insight from contracts. Specifically, dTrax will be used to standardize the intake of legal contracts, generate and edit contracts within a web interface, and identify and monitor changes in contracts. The proposed research project aims to improve dTrax by automating contract checking and analysis. This includes categorizing sections of a contract into clauses, determining whether the clauses satisfy predetermined requirements, and identifying risky clauses in contracts. Through this research, Deloitte will be able to bring tangible economic value to the legal sector, which can translate into improved decisions by companies trying to enter a legal agreement.

Developing Provisional Publishing Guidelines Informed by Indigenous Storytelling and Knowledge Keeping Protocols

This project will develop provisional guidelines for publishing texts for and by Indigenous people. Each story, each storyteller, and each nation presents a different case, and there can be no set of rules. The objective is to outline a provisional, adaptable set of principles, questions, and considerations to guide the process of selecting, editing, publishing, and marketing manuscripts. The knowledge base in this area will be expanded via interviews with professionals and community members with insight and experience to share; a comprehensive review of existing scholarly and other literature on the subject of Indigenous publishing; and through learning by doing as I work to republish several classic titles at Theytus Books. The guidelines will be a tool intended for continual adaptation by Indigenous authors, editors, and publishers, and will be submitted for review and feedback by my colleagues in the field of Indigenous publishing.

Constraining the age of deformation and alteration in the Eastern Highlands Shear Zone, Appalachian Mountains

The Eastern Highlands Shear Zone (EHSZ) is a zone of highly strained rocks located in the Cape Breton Island portion (Nova Scotia, Canada) of the Appalachian mountain belt. It is interpreted to separate rock packages of different ages and affinities. Recent field mapping of the EHSZ (this study) has identified deformation fabrics that are diagnostic of intermediate to shallow depths and a zone of associated hydrothermally-altered rocks that may be related to magma-induced poly-metallic mineralization along strike. The EHSZ, therefore, represents the ideal setting to examine how such structures can channel magma and other fluids into the Earth’s upper crust and promote the localization of ore deposits. The hypothesis leading this research project is that the continuous deformation of EHSZ through various crustal levels has provided the regional-scale “plumbing system” to focus magma-related mineralization along an orogen-scale fault that may extend from New Brunswick to Newfoundland.

Investigating Infinite Dimensional Models of Quantum Computation

As in the classical case, quantum computers are modeled by circuits: that is, basic components connected by wires. Circuit models of computation are used for optimization, minimizing the cost of implementing quantum computers. There are various circuit models for quantum computing with different features. Models of quantum computation generally do not allow for information to be copied or deleted. Similarly, models of quantum computation are typically time-reversible: meaning that circuits can be run both forward and backward. When a model of computation supports quantum teleportation, the wires of the circuits may be bent. Quantum teleportation and time-reversibility allow quantum channels to be modeled. These represent the combined transmission of quantum and classical information between systems. However, wires between infinite dimensional systems do not behave in quite the same manner, in particular, they cannot be bent. Circuit models which allow wire-bending are very useful and well-explored. I will investigate infinite dimensional models of quantum computation in which a more general variant of wire-bending can be performed. This, it is believed, allows the transport of ideas from the conventional circuits to the infinite dimensional case.

Establishment of a Pressure Ulcer Biobank from Spinal Cord Injury Patients

Pressure ulcers (PUs) are a major health problem in people with spinal cord injuries (SCIs). PUs increase hospital admissions, with 25% of SCI treatment cost linked to PU management. Currently, treatment is limited and new therapeutic options are required. Wound fluid from PUs is invaluable for researchers as it allows a detailed study of the disease. However, no system is in place to systematically collect and store SCI PU wound fluid in British Columbia. In response, the aim of this proposal is to collaborate with the Rick Hansen Institute (RHI) to establish a PU biobank. The RHI have access to large numbers of SCI patients and have developed a non-invasive technique to extract would fluid from PUs. Together, this biobank will provide an invaluable tool for researchers and assist with the overall goal of RHI to accelerate the translation of discoveries into improved treatments for people with SCIs.

Analysis of Nanofilter Fouling in Potable Water Treatment

The province of Manitoba has many lime/soda water treatment plants unable to meet the latest provincial quality standards for chlorine disinfection by-products – trihalomethanes (THMs). As a result, many municipalities are looking for other alternatives. One suitable alternative is a dual membrane plant using ultrafiltration and nanofiltration (NF). Typical surface water sources in Manitoba contain major membrane foulants in relatively high concentrations, specifically DOC and hardness. Control of membrane fouling allows for longer filtration runs and for less water to be rejected, which is directly related to the operating cost of the plant. Membrane fouling is influenced by the quality of local water, and the effect of the water of the Canadian Prairies on membrane filtration is not well understood. We propose to study the mechanism for NF membrane fouling in a Manitoba plant, and gain knowledge that will enable its control.

Dinosaur bonebed amber: paleoecological and paleoenvironmental reconstructions

Amber, or fossilized tree resin, is a natural trap that provides a valuable source of palaeontological data on insects in ancient ecosystems. The chemistry and composition of the amber itself can also be used to explore which trees produced each amber deposit, and what conditions these trees lived under. Western Canada contains multiple amber deposits near the end of the Cretaceous period (66 million years ago), and small quantities of amber have been found in the sediments of dinosaur bonebeds. The proposed project will sample amber from two dinosaur bonebeds in Saskatchewan (the quarry of “Scotty” the T. rex, near Eastend), and Alberta (the Pipestone Creek Pachyrhinosaurus Bonebed near Grande Prairie). This amber will be used to learn about conditions in the surrounding terrestrial habitats, and will help fill in a gap in the fossil record of insects that occurs near the end of the Cretaceous.

Visualization of Single and Multiphase Flow in Porous Media

Unconventional hydrocarbon require novel technologies to improve the recovery factor of the vast reserves in North America. An alternative to increase the reserves is enhanced oil recovery (EOR). EOR is a technique that reduce capillary effects and/or improve the mobility ratio by the injection of materials not normally present in the reservoir. This summer research project aims the mechanism understanding of multiphase flow in porous media using in situ visualization. The research activities include: (I) laboratory training, (II) laboratory-scale visualization of single and multiphase flow, (III) data collection and processing, and (IV) results analysis. The expected outcome will be a better understanding of the defined system in order to evaluate the potential results for a field test.

Axiomatizing complementarity in infinite dimensions

Quantum mechanics explains the behavior of particles in sub-atomic levels. A feature of quantum mechanics is that measurable properties of physical systems occur in complementary pairs : one can ‘know’ the value of only one of the paired properties at any point in time. For example, measuring the position observable changes the value of the momentum observable since they are complementary.

Recently, a mathematical framework has been developed in which one can describe many quantum phenomena including complementarity using simple diagrams. The diagrammatic framework is developed using Category Theory which is a branch of modern mathematics. The framework can currently describe only finite-dimensional quantum systems. However, there are important quantum systems which are infinite-dimensional such as quantum computers that compute over continuous variables. Hence, there is a practical need to extend the current diagrammatic framework to include infinite-dimensional systems and to describe quantum phenomena in the extended framework. In this project, we generalize the existing diagrammatic framework to describe infinite dimensions and attempt to formalize complementarity in this new framework. We will also apply the formalization to describe some of the existing infinite-dimensional quantum protocols.

User experience heuristics in extended reality

We now live in a world where people can communicate, in real-time, both directly and remotely (e.g., Skype); and where new technologies (e.g., virtual reality, robots, driverless cars) are blurring the traditional lines of human-human and human-computer interactions. The digital world of industry is experiencing a revolution. Products that have traditionally been displayed on computer screens, smart phones, and tablets as two-dimensional (2D) experiences are now beginning to be displayed in 3D, as augmented, mixed, or virtual reality. To what extent the heuristics regarding the delivery of effective 2D experiences applies also to 3D interactions is largely unknown. The aim of the present work is to address this issue.