Projets novateurs réalisés

Peer Information and Substance Use Decision Making in Street-Involved Youth

This research aims to learn more about the way that homeless and street-involved evaluate information regarding substance use from their peers. To this end, this research team will work with Operation Come Home, a non-profit organization supporting homeless and at-risk youth in downtown Ottawa. The research team will recruit and interview 40 – 50 youth at Operation Come Home to assess a number of factors relating the way that they evaluate peer information regarding drug use. The interviews will be analyzed to search for themes relating to the way that street-involved youth receive and evaluated substance use information from their peers, and whether the use of peer information is related to subjective indicators of well-being. The results of this research will be directly communicated to the primary care workers at the partner organization, in order to facilitate improvements in their communications of high-quality information regarding substance use to this population of youth.

Extension of AWSOM to Strut-Braced Wings

Bombardier is interested in investigation a novel aircraft configuration known as the strut-braced wing configuration, which is has the potential for improved fuel efficiency relative to current tube and wing aircraft. However, Bombardier’s preliminary multipdisciplinary design optimization tools need to be extended in order to be applicable to this configuration. The intern will extend the capabilities of Bombardier’s tools to enable sizing of strut-braced wing. This will allow accurate estimation of wing weight and stiffness properties of strut-based configurations, thereby facilitating design and evaluation of such configurations. This capability will assist Bombardier as it examines various options for future aircraft.

Combining Seq2Seq Models with Collaborative Filtering Techniques for Explainable Recommendation

Layer 6 builds state-of-the-art recommender systems for TD’s online businesses. Collaborative Filtering (CF) is a common recommendation approach that widely adopted by many e-commerce platforms. Modern CF algorithms attempt to exploit latent features to represent users and items, which can lead to the lack of transparency of the recommender systems. In order to build a trustworthy recommender system, it is necessary to provide explanations associated with each recommendation so that users can understand why a specific item has been suggested. The proposed research project would explore the potential of combining sequence-to-sequence (seq2seq) natural language generation models with collaborative filtering techniques into a multi-task learning setting. The result would be a recommender system that could predict customers’ needs with a high degree of accuracy, while producing effective, personalized explanations. Such explainability would build trust between the recommender system and TD’s customers, and accordingly drive sales and customer loyalty.

Categorisation des charges de proiet: outil d’aide a la decision

quebecoise de genie-conseil et de gestion de projet, qui exerce dans cinq secteurs que sont

I’energie, !’industrie, les transports, les infrastructures et Ie batiment. L’objet de cette etude est de

mettre a la disposition de la haute direction de cette organisation, un outil qui lui permette d’affecter

ses competences internes aux differents projets que realise I’entreprise.

L’interet de cet outil se situe a un double niveau. Premierement, au niveau de I’organisation, il

facilite la constitution rapide des equipes de projet par une affectation rapide et efficiente des

competences. Deuxiemement, au niveau des charges de projets eux-memes, r outil permet une

reconnaissance sociale, en donnant une visibilite sociale des professions.

En fin de compte, routil propose est un moyen de placer rentreprise dans une dynamique de

developpement.

An Artificial Agent for Light Switch

Ecobee is a home automation company that makes thermostats for residential and commercial use. Smart lights are the second-most desired home automation devices after thermostats, but being able to remotely control the lights is only a small part of the vision. ecobee light switches have embedded microphones, a speaker, and far-field voice technology that allows the user to control the lighting, thermostat, and other smart home products. One major goal of this project is to create an artificial agent, which monitors and controls the light switches. To make this concept a reality the agent will gradually learn the residents’ schedule and behavioral patterns, and make decisions accordingly. The light switch agent aims to estimate the probability for turn on and off actions by leveraging machine learning and AI, in a non-deterministic environment.

Examining Media for Fraud Detection

Nowadays a corporation’s public image plays a major role in that company’s decisions and financials. This project involves predicting fraud and errors within the financial statements of publicly traded companies. The goal is to incorporate information such as press releases and industry media coverages to provide an insight to these companies under audit and their industries. Ultimately, this would be used as a tool to assist auditors identify fraud and errors within these financial statements.
This project covers analyzing collected media information for features such as corporate sentiments, public opinions, and trend predictions. This can help identify misstatement within annual financial statements, stock manipulation, and other issues.

Fraud Prevention in Real-Time B2B Payments Using Streaming Algorithms

The Pungle payments-as-a-service platform delivers low cost, real-time, friction free business disbursements, peer-to-peer (P2P) transfers, and B2B supplier payments. Pungle’s mission is to enable businesses with a digital payments platform that provides real-time disbursements and transfers. The problem that arises with digitization of business payments is higher risk for fraud due to its electronic nature. Therefore, there’s a need to be absolutely certain that both the sender and recipient of payments are the intended parties and that there are no anomalies in payment volume and frequency. This project is to build a streaming data pipeline, including data warehousing, that will allow us to log and persist transaction data for both audit trails and as a data set with which we will then develop and train real-time fraud prevention system using the latest research in streaming / machine learning algorithms.

Privacy Enhanced Decentralized Identity System

Currently, some public and private organizations have implemented various identification verification solutions to manage identity authentication. The idea of using a third-party identity provider (IdP) to access a relying party (RP) is not new, and both RP and IdP have their benefits as they can only be connected once in a federated identity ecosystem. While the deployed identity brokerage system has provided participants with great utility, it was pointed out that the principles they designed had several security and privacy gaps. The potential weakness of the federated identity access system comes from a central observation point. It is clear that the federal identity ecosystem needs to evolve to meet the challenges described. In our research, the goal of the project is to establish a decentralized system to solve the identified problems.

Development of a Bidding simulation environment

Addictive Tech Corp is a fast-growing ad-tech company. They use real-time advertisement bidding software which is massive and sophisticated. The actual dynamics involved in any given bid are complex and hard to predict. This makes writing test logic for such a system cumbersome and catching all corner cases next to impossible. Because of the scale of operations, understanding the environment in which the bidding software operates is difficult. This is problematic as such software needs to be highly optimized to be competitive. Creating a simulation of the environment the system operates in allows for a controlled study of said system and environment. Having fine-grained control over environment parametrization in the simulation will make testing and analysis more streamlined.

Cognitive Risk Sensing Using Deep Learning

CRISP is an international Deloitte development initiative aimed at helping some of our largest clients understand and managed corporate risk. CRiSP stands for “Cognitive Risk Sensing”, and it centers around using large sources of mostly unstructured data (i.e. 10% sample of all of Twitter, thousands of news aggregators, etc.) to understand and forecast risk for the clients. The goal for the student is to apply new methods in machine learning, data mining and natural language processing to extract user opinion of products from social media and customer feedback. The challenge is that such data is very large in volume and may be sourced from multiple distinct domains.

Sex: It’s a matter of the heart

Heart failure is a complex cardiovascular disease with increasing global burden while the prognosis for patients remains poor. Risk factors and the type of heart failure differ between men and women. These differences can be due to sex – referring to biological differences – or gender – referring to social differences. In our project we will study the role of genetics in the different types of heart failure in men and women, using models that distinguish the contribution of both sex and gender. We will use genome-wide genotyping data from two AstraZeneca randomized clinical trials, the Montreal Heart Institute and the United Kingdom Biobanks, totaling >10,000 heart failure patients.
The partner organization, AstraZeneca, will benefit from a better understanding of the genetic, sex and gender specific determinants of heart failure which can lead to a better definition of patient subgroups driven by distinct mechanisms and may ultimately lead to new improved therapeutic interventions.

Les impacts des pratiques de gestion en santé et mieux-être au travail sur le profil de réclamations d’assurance des entreprises

Ce projet de recherche vise à mesurer le lien entre les pratiques de gestion en santé et mieux-être au travail, incluant l’utilisation de la technologie Dialogue, avec le profil de réclamations d’assurance des entreprises. Pour ce faire, un échantillon représentatif des 1 000 entreprises couvertes par SSQ Groupe Financier qui aura accepté de partager ses données en lien avec son profil de réclamations d’assurance sera d’abord sondé par questionnaire. Ce questionnaire visera à recueillir à la fois de l’information sur les pratiques de santé et mieux-être au travail et la capacité à changer des entreprises sondées. L’équipe cherchera ensuite à mettre en relation la maturité organisationnelle des entreprises sondées en termes de pratiques en santé et mieux-être avec le profil de réclamations d’assurance des entreprises, à savoir le taux de réclamations, les durées d’absence au travail (court terme ou long terme) et les coûts directs et indirects. TO BE CONT’D