Projets novateurs réalisés

Classification and Anomaly Detection of Network Traffic at the edge using Transfer Learning

Anomaly network traffic threat detection has been used in many companies to detect threats. Usually, it is deployed on-premises for efficiency and privacy reasons. Anomaly detection system usually consists of a machine learning algorithm that needs tons of time to train and test. Because each company may have different network setup, it is hard to build one system that suits for everyone. So, we need to customize the system for each client. Nevertheless, it is very inefficient to retrain the algorithm from zero for each company. Therefore, we hope to prove that by using transfer learning, we can move the knowledge from an existing detection system to a new network environment without retraining the whole algorithm.

A multi-cultural comparative study into the influence national level variations have on desistance from crime

The study of desistance – researching how and why people refrain from criminal behaviour – has gained increasing attention from both academics and policymakers in recent years in many countries throughout the world. Yet despite the growing strategic importance of more comprehensively understand the processes of desisting to reduce national reoffending rates, there remains a significant gap in empirical literature which comparatively explores the impact national-level practices and structural divergence has on desistance from crime. Therefore, this research project will examine key factors impacting individual’s desistance found in a variety of mono-cultural studies through a multi-cultural, comparative project of Scotland and Québec (a province of Canada). By doing so, this project will explore how national level characteristics and processes, societal institutins, cultural and religious beliefs and economic systems shape processes of desistance from crime. This will enable a more informed understanding of what national level variations support and inhibit individual’s ability to not refrain from future offending behaviour.

Water Quality and Waste-Solids Dewatering at a Zero-Exchange Recirculating Aquaculture System

To meet growing demand, the aquaculture industry is building increasingly larger land-based farms. There are few locations which can provide enough fresh water for their operation and the industry is forced to reduce its water requirements further by modifying recirculating aquaculture systems (RAS) to incorporate zero-exchange technology. Cooke Aquaculture Inc. is currently constructing a zero-exchange RAS at Johnson Lake, located in Pennfield, New Brunswick to produce salmon parrs. This facility will be the first of its kind in the Americas. This research focuses on two possible problems which may arise with the new system. The first problem is with regards to the possible accumulation of waste metabolites in the water. A water quality monitoring program will be implemented throughout the first production cycle, and any accumulations will be remediated as they begin to arise. A model to predict nitrogen and phosphorus concentrations in the waste stream based on feed loading and measured equipment efficiencies will also be developed for optimization purposes. The second problem deals with the dewatering of waste solids generated in the system. The effectiveness of a centrifuge will be evaluated, first using lab-scale tests to optimize polymer dosing, then applying the lab-scale findings to an industrial centrifuge.

Tetraquarks and Quantum Computing

We will perform a computational calculation to determine the properties of a proposed, currently unobserved type of particle called a tetraquark, within our current understanding of particle physics. This will allow experimentalists to search for such particles, and provide tests of our theory (Quantum Chromodynamics), which we currently find to be in agreement with observations. We will also investigate new techniques for performing similar calculations in future on a new type of computer, a quantum computer, which is currently in development and provides an exciting new way of performing calculations in future.

Impact of CO2 and salinity in aquaculture on physiology, growth and health of coho salmon

As of 2017 salmonid aquaculture was worth $22 billion USD per year with the UK contributing $1.4 billion USD and Canada responsible for $988 million USD. However, despite UN directives stating a need to double production by 2050, growth is hampered by negative public perception. Typically salmonid aquaculture combines land-based freshwater hatcheries with sea-pen rearing systems. While requiring lower maintenance costs the use of sea-pens increases risk of disease in farmed fish and has been linked with parasite overspill into wild populations of salmon causing serious declines in native populations. As such there is a demand to reduce the duration of the marine grow out phase, or transition entirely to land based farm systems (referred to as recirculating aquaculture systems – RAS) which largely avoid many of these problems. However, thus far development of RAS farms has been limited due to reduced growth observed in RAS compared to pens, and the scale of RAS required to maintain fish up to harvest size. Reduced growth and adverse health outcomes have largely been attributed to various issues relating to water chemistry (e.g. CO2, salinity, pH etc.).

Incorporating Statistical Analysis of the Extensive SAMPEX Dataset to Improve Space Weather Modelling

The Van Allen Radiation Belts are a very dynamic region of space, occupying the region between a few hundred to many tens of thousands of km above the Earth’s surface. Hundreds of commercial, scientific and military satellites occupy this region of space and are therefore vulnerable to the damaging effects of the highly energetic charged particles within it. Understanding, to the point of predictability via radiation belt models, the dynamics of the Van Allen belts, is essential for protecting future infrastructure, which increasingly relies on satellite-based technology. This project seeks to combine several new analysis techniques from both the UK and Canada, to shed new light on long-term spacecraft data. This will provide a platform for a deeper understanding of radiation belt dynamics over an extended period of time, improving current models and therefore increasing our capacity to predict future behaviour.

RECHERCHE CIBLÉE SUR LES ARCHITECTURES LOGICIELLES POUR CAMÉRANUMÉRIQUE ULTRA SENSIBLE

La nouvelle entreprise montréalaise Nüvü Caméras développe, manufacture et

commercialise des caméras reconnues comme les plus sensibles au monde. En plus de

quadrupler l’efficacité des télescopes en astronomie, la technologie EMCCD (Electron

Multiplying Charge Coupled Device) de Nüvü Caméras permet d’augmenter la

compréhension scientifique dans des domaines aussi importants que le diagnostic

biomédical, la pharmaceutique et la recherche sur le cancer.

Dans le cadre d’un stage MITACS, l’étudiant de l’École de Technologie Supérieure sera

exposé à contribuer au développement du logiciel embarqué et de contrôle d’une caméra

EMCCD. Cette réalisation devra s’appuyer sur une recherche de la meilleure architecture

permettant d’atteindre les objectifs à court, moyen et long termes, ce qui inclut la facilité à la

porter sur plusieurs plateformes. De la conception à la documentation, le stagiaire aura à

évoluer à travers tout le cycle de vie du logiciel tout en se familiarisant à la technologie

EMCCD canadienne.

Costs of fluoroquinolone resistance in clinical E. coli: a potential explanation for similarities in resistance between the UK and Canada

Antibiotic resistance is one of the biggest challenges faced by modern medicine. Countries operating public healthcare systems need to effectively treat patients while practising responsible antibiotic stewardship. The rising prevalence of resistance means, however, that many current first-line antibiotics are becoming ineffective. In particular, resistance to the the fluoroquinolones, a broad-spectrum class of antibiotics, is on the rise globally. Curiously, the prevalence of fluoroquinolone resistance is high on a national scale, despite differences in prescribing rates between countries such as the UK and Canada. The aim of this research is to understand if the high prevalence of fluoroquinolone resistance could be due to it having relatively little negative effect on bacterial growth in the absence of fluoroquinolone use. This would mean that resistant strains can remain prevalent even if fluoroquinolone use is reduced. To accomplish this, I will look at the effect of common resistance mechanisms in clinical strains of Escherichia coli, an organism where fluoroquinolone resistance is particularly common due to its use in treating urinary tract infections.

Graphene Oxide Langmuir Blodgett (GO-LB) Functional Thin Films

The aim of this project is to utilize the Langmuir Blodgett (LB) deposition technique at the University of Waterloo to form uniform thin films layers of PAni nanocomposites layers. Successful production of thin film PAni nanofibers will allow research into laser writing and the fabrication and testing of simple proof-of-principle flexible sensor devices. This will facilitate the development of a direct patterning method for producing high-resolution features, that is both simple and inexpensive, and, realization of a new class of polymer printed devices and soft robotics.

Wireless charging based on capacitive-inductive resonance

This project is about a novel chip-based wireless power transfer system. Its integrated and flexible design eliminates the need for power cables for any DC device in homes, offices, parkades and power generating facilities. Unlike existing solutions, the system operates simultaneously with multi electricity requirement devices, has extended range, generates no heat, can transfer power through solid material, while increasing the effective utilization of harvested electricity. It does so at a lower cost per power density than its competitors.

Biosensor for Detecting Pharmaceutical Residues in Water

The intention of this research project is to detect pharmaceutical agents in water systems, specifically carbamazepine, ciprofloxacin, and diclofenac. These substances present an issue as they compromise the water quality and have low removal rates from conventional water treatments. The use of nanotechnology in biosensors holds the promise of high sensitivity and miniaturization. Here the detection of these agents is based on a single DNA biochip for the parallelized analysis of the mobility of DNA tethered particles. A drug molecule is grafted to the DNAs at a precise location and the recognition element is coupled to the particles. In a competition assay, the particles initially linked to the DNA are released after the binding of a free drug to the recognition element. The binding event leading to an increase of the amplitude of motion of the particle is monitored by videoimaging. The biosensor would be tested in a synthetic water solution similar to river water with varying concentrations of each drug. DNA molecule biosensors have a wide range of applications for future water monitoring, and is an area of research that must be further explored.

Ships arrival times prediction using deep learning

The objective of this research project is to use Deep Learning in predicting ships’ arrival times of a medium size multi-commodity bulk port. The outcome of this project will serve dynamically in updating berths schedules to minimize ports turnaround time and build models closer to reality. We will be using ships trajectories data or AIS data combined with some exogenous data such as weather data, physical structure of ships, Machine conditions, unforeseen events and seasonality. Also, we will focus more on the real time estimation of arrival times by building flexible model for short-term and long-term prediction horizons. A first objective will be to prepare properly the data that will be used in building the predictive model, as the problem may arise in real-time data management in terms of computing and storage resources. Also, the use of multiple data sources will require a specific treatment related to data handling and feature engineering. The second objective will be to build models based on Artificial Neural Networks, train them and select the best one. Then, the last objective will be to deploy the chosen model in order to make prediction on production data.