Projets novateurs réalisés

Assessing the role of climate and lithology in vegetation distribution

The proposed project will use machine learning to predict vegetation biomass across the state of California using climate variables (precipitation, temperature). This will create a map of what vegetation is expected to be in a given climate regime. This modeled vegetation distribution will be compared to a map of observed vegetation to identify regions where climate cannot explain the observed patterns. In these regions, we hypothesize the subsurface bedrock exerts a control on plant productivity. We will compare with geologic maps to examine how rock type may correspond to areas of limited vegetation. The expected outcome of the project is a map of where climate data can and cannot predict plant biomass, as well as statistical relationships between each of the input variables and plant biomass. These results will inform future studies on how the subsurface controls water availability to plants.

A comparison of heart rate variability and brain activation during emotional regulation in athletic groups

Heart-rate variability (HRV) may be used to index important aspects of brain control of the heart during emotional regulation. Yoga may enhance athletic performance, in part, by facilitating emotional regulation. Lululemon Athletica is interested in developing biofeedback devices designed to further enhance yoga related emotional regulation and for use to educate both their in-store work-force (i.e., educators) and clients (i.e., guests). Prior to the implementation of such techniques, the relationship between brain activation and HRV must be established. HRV has been shown to be a promising biofeedback measure during sport performance. Functional magnetic resonance imaging (fMRI) measures task related brain activation, allowing the direct relationship between brain activation and HRV during emotional regulation to be evaluated. The Vancouver based company, Lululemon, and the Brain Behavior Lab (BBL) at the University of British Columbia will collaborate to test the effect of yoga practice on the relationship between HRV and brain activation through a combined HRV-fMRI approach. A group of expert yoga practitioners, recreational yoga practitioners, and recreational non-yoga athletes will be compared to determine whether yoga expertise enables higher HRV and related to patterns of brain activity during varied emotional stimuli. TOBECONT’D

Leading Indicator Generation from Cryptocurrency to Detect Fraud

The goal is to understand the behavior of cryptocurrency chains when fraud amounts are transferred from the MasterCard network to cryptocurrency exchanges. Based on this understanding, we need to come up with an Index score or Leading indicators of certain pattern that indicates high fraudulent activity in the MasterCard network. We need to build such indicators at least aggregate level and verify the importance of the same in predicting future fraud. The algorithm should be able to continue to come up with new patterns and anomalous activity for further analysis.

Aircraft systems noise / Bruit de systèmes avion

De multiples sources vibrantes intégrées dans les avions sont à l’origine de niveaux acoustiques gênants à l’intérieur de la cabine de l’appareil. Ce bruit, appelé bruit de structure (SBN, acronyme de l’anglais Structure Borne Noise), résulte de la transmission des vibrations internes des sources à leur support et est un élément majeur dans la conception des avions puisqu’il affecte le confort des passagers. Des méthodes d’estimation du SBN basée sur des quantités dynamiques intrinsèques aux sous-structures découplées existent et sont nécessaires pour réduire ce bruit durant la phase de conception. Cependant ces méthodes restent pour le moment appliquées principalement sur des structures en laboratoire leur capacité à prédire précisément le SBN dans un contexte aéronautique reste inconnu. Ce projet vise à établir les meilleures pratiques de caractérisation du comportement dynamique de sources vibrantes aéronautiques typiques dans un environnement industriel dans le but d’estimer le SBN qu’elles induiraient.

AI-driven multi-target neuroprosthetic optimization

We will design novel BO strategies to parallelize multi-target neurostimulation optimization.
The development of large-channel-count neural interfaces aims at obtaining high-throughput communication with the nervous system by parallelizing the flow of information over multiple subregions. For instance, the motor cortex, spinal cord and nerves host representations of multiple bodily movements within highly-interconnected topographies. Neural interfaces are currently unable to explore these arrangements to evoke diverse functional motor outputs. This mapping is restricted to the human capacity to focus on one target at a time. This technological gap severely limits the development of multivariate neuroscientific investigation and multimodal neuromodulation intervention.
We propose a new algorithm for dynamical multi-target Bayesian optimization, suitable to parallelize multi-target optimization, allowing full identification of the controllability space (i.e., all possible outputs that can be generated with the available inputs) of a single interface.

Comment les méthodologies de design thinking peuvent supporter l’amélioration continue des processus en commerce de détail?

Le projet de recherche est une expérimentation visant à évaluer l’intérêt de la mobilisation de la méthode du design thinking pour l’amélioration des différents processus managériaux et commerciaux d’un centre jardin. À l’aide du la boîte à outils et de la méthode du design thinking, les différents processus seront analysés. Par la suite, des recommandations seront formulées dans le but de les améliorer et de les mettre à jour toujours en suivant les principes du design thinking. Une évaluation critique de la démarche sera développée.
L’organisation partenaire tirera des recommandations spécifiques à sa situation pour l’amélioration de ses processus et sera initiée à l’approche du design thinking. Les autres commerces de détails pourront s’inspirer de ce projet de recherche pour mobiliser le design thinking dans leurs démarches d’amélioration.

Lolli 37 Designs – Junior Staff Accountant and Social Media Marketing Project – Jasmine & Neha

Loulou Lollipop is a Canadian company founded by the twin sisters Eleanor and Angel. In pursuit of better, more attractive, and safer products for every newborn adventure, we are redefining available baby products. We are one of the fastest-growing, female-led, apparels and accessories brands for children in North America. Our products are designed in Vancouver and created from the heart. We are a team with imaginative style, uncompromising safety standards, and a desire to create a better world for future generations. This internship project, led by two KPU interns, will bring innovation to both our financial practices and our digital marketing strategies. Our Junior Accountant Intern will revamp current systems, eliminate accounting inefficiencies, and collaborate with our controller to start a manual for best financial practices to be followed moving on. The Social Media Marketing Intern will work on creating a strategic plan for all of our digital media efforts, including our social media platforms, and will analyze best usage. As an active participant in data collection and analysis, the intern will consistently improve the company’s digital marketing practices, identify gaps, and provide remedies.

Évaluation des services écosystémiques associés à l’hydrologie et à la qualité de l’eau à l’échelle des bassins versants de la région d’Edmundston selon l’approche du capital Nature

La municipalité d’Edmundston souhaite se doter d’un plan de gestion des actifs naturels pour la mise en place d’une gestion intégrée de l’eau et son adaptation au changement climatique. Pour ce faire, elle vise à documenter de manière scientifique, explicite et cartographique les rôles des écosystèmes naturels (milieux humides, peuplements forestiers, bandes tampons riveraines) pour l’hydrologie et la qualité de l’eau à l’échelle des bassins versants. Plus spécifiquement, ce projet permettra d’évaluer les services écosystémiques rendus par les actifs naturels dans les conditions climatiques actuelles et futures, d’établir des scénarios d’aménagement du territoire et d’investissement dans la protection et la restauration de milieux naturels, et d’optimiser des stratégies pour le maintien, voire l’accroissement de la rétention des sédiments et la purification des eaux de surface, l’atténuation du risque d’inondation et le maintien de l’approvisionnement en eau potable durant les sécheresses estivales, malgré les impacts des changements climatiques. Ainsi, ce projet pourra contribuer à un meilleur aménagement des bassins versants et à une atténuation des impacts climatiques sur la ressource en eau.

Geometry recognition from computer-aided design drawing files

Currently, the service provided by Axya links manufacturers and buyers of manufactured goods. To find a perfect match between a manufacturer and a buyer, the files they share need to be analyzed. CAD engineering drawings represent most of the files shared. However, manual analysis of large amounts of engineering drawings is too time-consuming while requiring high engineering expertise. Based on this, it is proposed 1) to investigate ways of automating the analysis of engineering drawings with computer vision, 2) to retrieve and analyze crucial geometric features from engineering drawings, and 3) to reconstruct 3D geometry from 2D data in engineering drawings.

Valorisation des données dans les services d’urgences

Les patients dans les services d’urgences ont besoin d’accéder aux soins dans des délais spécifiques afin d’assurer le bon traitement au bon moment. Le nombre limité des ressources humaines et la mauvaise planification ne permet pas l’exploitation optimale en urgence. Le déséquilibre entre la demande d’accès aux soins et la capacité des services d’urgences devient de plus en plus important surtout avec la croissance continue de la population, ce qui influe sur la satisfaction des patients et les dépenses médicales telles que le temps supplémentaire travaillé par le personnel. Les gestionnaires des services d’urgences font face aux longues durées d’attente des patients et un budget médical limité. Donc, ils sont obligés de faire une gestion efficace des ressources en réduisant les coûts et en assurant une bonne qualité de service.

Ce projet vise essentiellement à améliorer le processus de gestion en urgence en minimisant le temps d’attentes des patients en valorisant les données opérationnelles, et en utilisant l’intelligence artificielle.

Active noise cancellation of various noisy environments, especially construction sites, using advanced deep learning.

There are many commercialized active noise cancellation (ANC) headphones and headsets that try to reduce unwanted background noise. You can simply listen to your favorite music or talking with your loved ones in crowed areas using them without any distractions. The overall working principle of them is to generate an inverse signal to unwanted noise to neutralize it inside the ear drum. These devices are specially produced for single users. But how about cancelling excessive noise in car passengers’ cabin or working places that loud machineries are working? In this research, we are going to build an ANC device suitable for open space environment to cancel any kind of undesirable noise while keeping human voice unchanged using advanced AI algorithms. The developed device can be used in factories, construction sites, libraries, hospitals and residential buildings in densely populated downtown areas.

Data Justice for the New Data: Contextualizing Data Storage and Protection in Decentralized Trust Technologies Used in Conflict Contexts

The goal of this research project is to understand how well blockchain and distributed ledger projects conducted in humanitarian settings live up to their potential for benefitting technology users. The research will consider this question through the lens of data justice, an emerging framework that looks at the storage and protection of digital data and how technology users are treated when they produce digital data. This framework is relevant to the production of the data generated by the way individuals lead their social and economic lives online. The project partners with Peer Social, an organization currently building a solution for displaced individuals in Ukraine to use this type of data to attest to their identity and ownership of homes, lands, and properties (HLP) that may have been destroyed in the war caused by Russian aggression. The project will deliver a set of recommendations for Peer Social and other technology providers working on humanitarian blockchain projects for how they might attend to data justice in their products and a scholarly research paper on the topic.