Projets novateurs réalisés

Functional MRI investigation of the neural processes underlying pain modulation in human participants

Cannabis has been known to treat ailments for thousands of years, including conditions such as pain, cancer, arthritis, glaucoma, multiple sclerosis (MS) and amyotrophic lateral sclerosis (ALS). Despite an increase in research dedicated to molecular and behavioural effects of cannabinoids, there has been no direct evidence to elucidate the effect of cannabis on pain-related areas in the human central nervous system. Using functional MRI we will identify how neural processes involved in pain are altered by cannabis in the brain, brainstem and spinal cord in healthy participants, and also how these processes are altered by fibromyalgia syndrome (FMS), a prevalent chronic pain disorder. This project will be an essential first step toward future research into cannabis-based treatment of acute and chronic pain for a variety of conditions including FMS.

Assessment of the mechanical properties on PVD coatings using scratch and indentation tests.

In manufacturing industry many tools are protected by coatings to increase tool life and productivity. It is only logical that mechanical properties of the coating will affect its performance. Therefore, it is essential to characterise the coating and understand its properties to tailor it for a specific application. These properties might be assessed with different characterisation methods such as scratch test and indentation. The objective of this study is to provide coating characterization for the partner organization and develop a standard procedure to assess coating properties. This will help the partner organization to optimise and develop coatings and keep track of quality.

An investigation into the operation and operational benefits of a new converter technology for supercapacitor charging

Batteries are main storage systems in many applications such as electric vehicles, shipping, transportation, and utility backup power. With the recent breakthrough in the supercapacitor technology, it is predicted that supercapacitors will challenge the batteries in many of these applications since their power delivery is much faster than the batteries. The current chargers are designed based on the requirements of the batteries. Considering supercapacitors as the prospective next generation energy storage systems, it is claimed that the deflection conversion technology offers much faster and significantly more efficient charging scheme compared to the current chargers. As an important consequence of this development, with the adoption of the new generation of supercapacitors in electric vehicles, drivers can charge the vehicles’ batteries on the go leading to saving time, reduced costs and benefitting the environment. This research proposal is intended to validate the superior characteristics of the chargers using the deflection conversion technology.

Tunable Fiber Laser Source based on Supercontinuum Generation

The objective of the project is to conceive and build a prototype of a tunable “white fiber

laser” i.e. a broadband light source. As it is based on supercontinuum generation in a

microstructured fiber, the source is expected to be compact and find applications in several

domains, from fundamental studies to biomedical imaging and spectroscopy as well as

sensing. Current supercontinuum-based tunable sources use a straightforward design where

a tunable filter is used to select a narrow band of wavelengths from the broadband source,

discarding most of the light, thus being very inefficient. The project consists in conceiving and

building a more efficient tunable laser source. The idea is to extract a narrow band of

wavelength from the supercontinuum, and recycle the remaining supercontinuum in the same

fiber in a cavity. As a majority of the light will be recycled rather than discarded, the process

is expected to be more efficient than the current scheme.

Deep-learning-based Fine-grained Furniture Classification and Winning Strategy Recommendation

The project aims to develop a novel deep learning based computer vision system to identify different categories and sub-categories of the furniture and the associated attributes (such as color, shape, style, and material). It will also develop an automated recommendation system that can learn from the massive historical data and the on-going stream of data to adaptively adjust the parameter combination for each product to maximize the chance of winning the competition against other companies. The competitive advantage gained by the new technologies developed through this project will help the partner organization, Cymax, to further grow and expand its business.

MalChain: Run-time detection of malicious virtualized components in Service Function Chains using machine learning

Intern will develop new methodologies based on system level logs to detect malicious activities in virtualized environments. To achieve this goal, intern must review the state-of-the-art and proposes new solutions to overcome the shortcomings of currently existing methodologies and introduce new functionalities. The proposed solutions need to be validated and published based on experimental results. The Research Department at the Ericsson Montreal location conducts research and development in cybersecurity, software defined networks and network function virtualization. The interns will work with a team of researchers towards accomplishing the aforementioned tasks through his/her expertise in platform security, machine learning, and cloud environment. TO BE CONT’D

Chain Certs: Development of a platform for organizational authenticity certificates creation

Blockchain is a decentralized and immutable data structure. The information stored on blockchain is tamper-resistant, immutable and transparent. Blockchain is an interesting platform for managing digital certificates without a central authority. Because paper certificates can be easily faked or tampered with modern computer skills. Additionally, using a central authority for issuing distributing certificates is inefficient.

In this project, we will analyze the security and scalability of different approaches to certificate management solutions using blockchain. This analysis will provide guidelines for certificate management in permissioned blockchains. The guidelines will be developed through the construction of a reusable and configurable testbed for blockchain performance testing and analysis of the results for the new configurations that exist within the certificate management space.

Détermination des facteurs limitants de la performance du triathlète

L’objectif de ce projet sera de mettre en place un protocole d’évaluation précis pour les triathlètes. Cette évaluation déterminera les points forts et les points à améliorer dans chacune des trois disciplines, c’est-à-dire en natation, en cyclisme et en course à pied. De cette façon, nous pourrons fournir les données nécessaires aux entraîneurs et aux athlètes afin que ceux-ci puissent obtenir un portrait spécifique des capacités de chacun de ces athlètes. De cette façon, il sera possible de quantifier les charges d’entraînement pour chaque discipline de sorte que l’aspect à travailler par l’athlète soit bien déterminé. Donc, cette évaluation sera utile pour orienter la planification d’entraînement des athlètes, effectuer un suivi précis ainsi qu’optimiser la progression de ceux-ci tout en améliorant leurs performances.

Determining position, speed and stride length using machine learning with sensor fusion based on ultrawideband local positioning system technology

Sensors that track human movement are becoming more and more popular in all kinds of applications including healthcare, sport and general human movement. However, traditional sensors generally have problems tracking individuals indoors and they are not very accurate when measuring subtle movements. Using innovative technology, new wearable sensors have been developed to track human movement that have solved the problems associated with previous sensors. Further development of these new sensors is still required and that is the overall aim of this project.

The goal of this project is to develop software that accurately calculates the speed and stride rate of athletes who wear a small sensor when they walk or run. To do this, we will compare the information we receive from the sensors with data that we collect in a laboratory using a video-based motion capture system, which is highly accurate. We will also use some advanced Artificial Intelligence techniques to process the information to help us develop the software. The newly developed software will allow the partner company to market and sell a new system that is very accurate and can be used indoors, giving them a major advantage in the marketplace

Plant level implementation of a model for real time tracking of composition changes to steel, slag and inclusions during ladle processing

The Ladle Metallurgy Furnace is used for adjustment of chemical composition and temperature, and control of tiny particles called “inclusions”. Controlling inclusions is carried out by adding calcium to modify the solid alumina or magnesium aluminate inclusions to less harmful liquid inclusions.
During ladle process, reaction of top slag, steel and inclusions occur simultaneously. Therefore, establishing a model to describe ladle process is indeed a challenge. The author developed a model to predict the chemical composition changes in molten steel, slag, and evolution of inclusions in the ladle during Ca treatment. The result of calculations was found to agree well with industrial heat data. However, the model is not in a form that can be used as a real-time tool. The overall objective of the current project is to develop a version of the model that will be sufficiently fast to use for real-time processing. Secondary objective is to calibrate the model for the ladle used in the KOBM stream at ArcelorMittal Dofasco

Modification éco-responsable de mousse à base de cellulose

es dernières années, les objectifs imposés par la gouvernance environnementale internationale ont conduit au développement de matériaux plus durables. Au Canada, cela a permis l’essor de nouveaux produits technologiques issus de l’industrie agroforestière et en particulier la valorisation de la cellulose pour le développement de nouveaux matériaux ‘vert’ à faible impact environnemental. Parmi les différents exemples on peut citer l’utilisation de la nano-cellulose cristalline dans le renforcement mécanique des matériaux composites. Récemment, l’utilisation des suspensions de cellulose cristalline ont conduit à la synthèse de mousses nanoporeuses. Ces nouveaux substrats biodégradables possèdent une porosité nanométrique contrôlée ce qui permettrai leur utilisation dans la filtration sélective d’agent polluants. Aujourd’hui, leur emploi reste limité à cause de leur forte hygroscopie (forte sensibilité d’un substrat à l’eau et à l’humidité). TO BE CONT’D

Optimization of group equivariant convolutional networks

The explosion of popularity of deep learning owes a lot to the success of convolutional neural networks, widely used in diverse fields including computer vision and natural language processing. Recently, the group equivariant convolutional neural network (G-CNN) was introduced, where equivariance of symmetries inherent in the data set is built in the architecture of the networks. While the G-CNNs has proven to exploit inherent symmetries more effectively than traditional CNNs, their architectural design and implementation require a deeper understanding of the mathematical concept of symmetries. We propose to develop better mathematical tools suitable for deep learning on G-CNNs, with two main goals: (1) improving the optimization methods of G-CNNs by exploiting the geometry underlying the inherent symmetry of the data set; (2) generalizing the architecture of G-CNNs to adapt to other practical learning problems, such as speech recognition and image processing.