Innovative Projects Realized

Coalescing around gender inequity frames on Twitter: The use of online campaign to #ElectMoreWomen in the Toronto 2018 Municipal elections

Using an advanced social media analysis tool developed by experts at Nexalogy, we will look at a Twitter campaign that focuses on gender-identity and inequality in an effort to challenge and change unequal representation of marginalized people (in this case, gender) in elected office and spaces of traditional political power. By looking at female-identifying challenger candidates who use this framing, and employ hashtags like #electmorewomen, we will compare and interpret what impact this has on a candidates level of engagement, perceived legitimacy, and also how the public accepts, reacts and challenges these framings of candidates.

Electrochemical Fischer-Tropsch Synthesis of Renewable Liquid Fuels from CO2

Despite a rapid decline of electricity costs, there is still demand for energy-dense liquid fuels, such as in heavy freight and air transportation. Liquid fuels can be synthesized from a mixture of carbon monoxide and hydrogen called synthesis gas (syngas). However, this process requires high temperatures and pressures, and is itself responsible for significant greenhouse gas emissions. We propose the use of electrocatalysis to produce these liquid fuels. To accomplish this, we will use computational modeling and machine learning methods to design electrocatalysts that efficiently convert CO2 or syngas into dense chemical fuels. These computational efforts will be validated through a parallel experimental approach that includes the fabrication of new catalyst formulations and the construction of prototype electrochemical flow cells.
This project will enable the synthesis of clean, energy-dense liquid fuels that can replace the use of fossil-derived fuels in industry and transportation sectors.

Accessible Data Platform for Dynamic experience study of Lifestyle Underwriting

We seek to replace or enhance the traditional underwriting approach (namely identification of insureds via a pre-defined fixed set of risk criteria) with one based on a set of dynamic protocols that are responsive to human behavioral factors for continual health improvement. We seek to provide a live and interactive in-market research dataset that can be used to explore the benefit of and improve data-driven approaches (namely artificial intelligence or AI) for immediate use in life & health insurance product development and actuarial risk assessment.

Development of improved power quality detection methods suitable for modern applications

Discontinuities of service, variations in voltage magnitude, and distortions in AC voltage waveforms constitute the different aspects poor power quality. A poor quality of power supply can cause malfunction of sensitive equipment and interrupt industrial processes, resulting in significant economic losses. Utilities and consumers are taking actions to maintain the power quality set by the standards. Monitoring of power quality at all levels in the power system is necessary to ensure adherence to standards, but specialized power quality monitoring equipment are expensive. Cost of monitoring can be reduced if monitoring functions are integrated to multifunction devices such as fault recorders or protection relays. However, most advanced power quality event detection methods require significant computing power and their implementation on multifunction devices is challenging. The proposed research aims to develop improved power quality detection methods suitable to implement on a resource constrained computing environment.

Patient Privacy Preservation through Federation or Encryption? A Comparative review and prototypes

The recent advances in machine learning based on deep neural networks, coupled with the availability of phenomenal storage capacity, are transforming the industrial landscape. However, these novel machine learning approaches are known to be data hungry, as they need to tune a huge number of parameters in order to perform well. As more and more AI based applications are being deployed to learn from personal data, privacy concerns are rising, and more specifically on sensible domains like medicine, finance or mobile related data. With the ubiquitous availability of cloud-based solutions at a very low price, privacy has now become even more sensitive. Moreover, privacy concerns seem to be two sided, as service providers would like to keep their models and learned weights private.
This research will focus on studying available solutions for privacy preservation in the context of medical data, and more specifically on volumes obtained from CT scans.

Étude de la dualité onde-particule sur une corde vibrante

Le sujet proposé porte sur la dualité onde-particule. On propose ici à l’étudiant l’étude expérimentale d’un système modèle qui vise à tenter une représentation mécanique du système quantique onde-particule. L’étudiant aura en charge un banc expérimental composé d’une masselotte pouvant coulisser sur une corde vibrante, le tout étant instrumenté notamment par un système d’imagerie rapide. L’étudiant sera en charge de faire varier les paramètres de l’expérience et de tenter de les interpréter en relation étroite avec la mécanique quantique et en particulier avec la théorie de De Broglie-Bohm.

Le stagiaire aura pour mission de partager son temps entre une analyse théorique du problème et d’une méthode expérimentale. Des simulations numériques (Python) du système ont déjà été développées dans l’équipe. Une prise en main et une analyse de ces simulations devrait également faire partie des activités quotidiennes du stagiaire.

Machine learning approaches for event prediction, relation modeling, and inference

Machine learning approaches are transforming fields such as finance, healthcare, electronic commerce, social networks, and natural disaster forecasting. We propose collaborative research that develops novel methods and applications of machine learning techniques for event prediction, modeling relations between entities, and inference techniques that can impact these domains. In the context of event prediction, we will develop methods based on the point process framework. We will develop novel models for learning the temporal distribution of human activities in streaming data (e.g., videos and person trajectories). Methods based on an integrated framework of neural networks and temporal point processes will be considered. For the problem of modeling relationships, we will build relational representations of entities, given graph structures describing potential interactions. Both supervised and unsupervised learning paradigms can be potentially utilized. Finally, we consider inference techniques for structured random variable spaces using deep learning approaches. TO BE CONT’D

Second Language Impact on First Language Grammatical Processing in Chinese-English Bilinguals

Most of the world speaks two or more languages and the first language clearly plays a role in learning the second, as we hear in foreign accents. But bilingualism research is now giving increased attention to effects in the opposite direction, where the second language impacts the first. This process is known as first language attrition. In 2017, Canadian researchers found the first neurocognitive evidence that immigrants begin to lose their native grammar and process their native language like their second language. However, we still do not understand how attrition occurs, and my PhD project focuses on characterizing the underlying mechanisms of attrition by studying Chinese-English bilinguals using state-of-the-art electroencephalography (EEG) methods. EEG can measure real-time language processing, and is sensitive enough to detect small processing differences. To distinguish attrition from other effects of bilingualism, I will compare different populations of Chinese-English bilinguals to monolingual Chinese controls. To this end, I am forging a new collaboration with Nanjing Normal University in China. Understanding the mechanisms of attrition is relevant for all bilinguals, and has important implications for treating language disorders in immigrant populations and language learning policy.

Rational Convexity in Complex Euclidean Space

The proposed project is focused on a research area of great interest that my supervisor, Dr. Rasul Shafikov, and Dr. Alexander Sukhov of the University of Lille, have devised in a number of coauthored papers published during the past few years. The research investigates the polynomial and rational convexity of compact sets in Euclidean complex space, notions that are of crucial importance in the general theory of approximation of continuous functions, uncovering deep connections to topology, Banach algebras, symplectic geometry, and other areas of mathematics. The objective of the project is to build on the existing work done by Shafikov, Sukhov and other researchers in the field, by giving a concrete answer to a still open problem that follows from their work. As a first step toward that goal, we seek to provide a criterion for the rational convexity of compact subsets of Euclidean complex space, with finitely many singular points of a certain type: hyperbolic complex points.

Trade Promotion Forecasting and Optimization

This project addresses two specific challenges related to promotional planning in the consumer-packaged goods sector. The first output will be to develop and evaluate models that statistically predict the impact of Unilever promotions on category and product share across different retailers. This effort will lead to the creation of trade promotion optimization techniques that enables planning of promotional activities. The second output is to simulate potential outcomes of various combinations of promotional programs across different retailers in such a way as to inform investment decisions early in the planning process. Finally, the project aims to develop a systematic approach for allocation of promotional spend that permits continuous adjustment as market and competitor conditions change.

Bicycle Aerodynamic Sensor Development

The interns in this project will help to develop a sensor system capable of measuring the aerodynamic forces on a bicyclist. The measurement will be done in real-time; i.e. the rider will be able to see as they ride down the road how changes in their position or equipment affect their drag, and hence be better able to tune their riding habits to go faster.

Scaling 3D Microtissue Production Using a Microwell System

Tissue engineering works to replace damaged tissue and organs, and has applications in treating diseases such as diabetes. To improve the performance of tissue engineering treatment and research, our lab has produced the microwell system which creates microtissues in the form of cellular balls. The microwell system has been used used internationally under the brand name AggreWell™, supplied by STEMCELL Technologies and allows researchers to produce enough microtissues to treat diseases in animal models, such as mice. This project will focus on scaling up the existing microwell system so that it can produce enough microtissues to clinically treat humans. The proposed designs for the microwell system must be validated and prototyped. Should a final functional design be identified, this research may lead to the production of a new tissue engineering tool which will be commercially pursued by STEMCELL Technologies in collaboration with the University of Calgary.