Projets novateurs réalisés

Explorez des milliers de projets réussis issus de la collaboration entre organisations et talents postsecondaires.

31 620 projets complétés

2978
AB
5221
C.-B.
856
MB
696
NL
899
SK
9419
ON
9858
QC
98
PE
619
NB
1192
NS

Projets par catégorie

Deep Learning Classification and model compression

Deep neural networks (DNNs) have achieved great success in many visual recognition tasks. However, existing state of the art deep neural network models are computationally expensive and memory intensive, hindering their deployment in devices with low memory resources or in applications with strict latency requirements. Therefore, a natural thought is to perform model compression and acceleration in deep networks without significantly decreasing the model performance. Developing models for inference on clients has multiple economical benefits, but it becomes difficult to match the performance of bigger architectures by simply training smaller architectures. Therefore, we have to look for solutions like Knowledge Distillation, Network Pruning, Quantization to obtain highly efficient models that can match the performance of bigger architectures.

Voir la description complète du projet
Superviseur du corps professoral :

Ioannis Mitliagkas

Étudiant :

Partenaire :

Jumio

Discipline :

Computer science

Secteur :

Professional, scientific and technical services

Université :

Université de Montréal

Programme :

Accelerate

Social and Community Economic Development Planning in Toronto

“Social and Community Economic Development Planning” is a BSI project is comprised of four internships associated with four non-profit partner organizations engaged in organizing, advocacy and planning for affordable housing and community supports in low-income, racialized and marginalized neighborhoods. The shared context for this work is the ongoing gentrification in downtown Toronto , coupled with displacement pressures associated with transit-led redevelopment in suburban regions of the GTA. Two projects entail research that will support the partner in developing community benefits strategies associated with major infrastructure redevelopments. Two focus more explicitly on researching possibilities for supporting the development of community land trusts that could take land off the market in perpetuity and preserve affordability for low-income residents and businesses as well as social enterprises. The partner organizations, who are among Canada’s leading practitioners of community economic development, will benefit from background research to support strategic planning, popular education and engagement of low-income and marginalized constituencies.

Voir la description complète du projet
Superviseur du corps professoral :

Katharine Rankin;Lindsay Stephens;Jason Spicer

Étudiant :

Partenaire :

Black Urbanism Toronto;North York Community House;Jane/Finch Centre;The Neighbourhood Land Trust

Discipline :

Sociology

Secteur :

Health and Related Sciences & Technology

Université :

University of Toronto

Programme :

Business Strategy Internship

Advancement of biosensor technologies for use in healthcare/cancer research and detection

The detection of specific carbohydrates is critical for several processes including biofuel production, textile finishing, food production and human health. These carbohydrates play roles in human health including stem cell differentiation, genetic diseases, and viral infection (i.e., COVID-19). Presently, the largest barrier to understanding the role carbohydrates play in human health is the lack of suitable tools to study them. Biosensors are devices that are used to detect and quantify the concentration of biomolecules or microorganisms. Protein-based biosensors, like those described in this proposal, are composed of one or more fluorescently labelled proteins that interact with a target molecule in solution. The interaction of the target molecule with the biosensor induces a change in the protein, which in turn, alters the position/ environment of the fluorescent dye. This change results in an altered fluorescence output, which can be measured. The challenge for developing new protein-based biosensors is that often multiple rounds of the design, build, and test cycles are required to find labeling positions that provide a change in fluorescence upon binding the target molecule yet do not disrupt the binding properties of the protein.

Voir la description complète du projet
Superviseur du corps professoral :

Trushar Patel

Étudiant :

Partenaire :

Allos Bioscience Ltd.

Discipline :

Life Sciences

Secteur :

Professional, scientific and technical services

Université :

University of Lethbridge

Programme :

Accelerate

Mimicking medial-knee brace gait using PCA and biofeedback

Knee osteoarthritis is a disabling disease affecting millions of people worldwide. Knee braces have been adopted as a treatment strategy to help manage osteoarthritis pain. These braces apply a force to the knee, which, in theory, reduces the joint loads in the knee. They also however change the way people walk, which can also reduce pain. The purpose of this research is to determine whether the positive effects of the knee braces are due to the altered walking patterns, or the force they apply to the knee. If the altered walking patterns are what reduce knee pain, patients with osteoarthritis can potentially reduce treatment cost and avoid wearing a brace. This research is being performed in collaboration with HAS-Motion, who has recently released a product required to perform this research. This project will test the capabilities of this new product, before it is distributed, in a real laboratory environment.

Voir la description complète du projet
Superviseur du corps professoral :

Kevin Deluzio

Étudiant :

Partenaire :

HAS Motion

Discipline :

Engineering

Secteur :

Biotechnology; Health and Related Sciences & Technology; Biotechnology; Health and Related Sciences & Technology

Université :

Queen's University

Programme :

Accelerate

Heart and Stroke Atlantic Canada Cardiac Arrest Policy Strategic Plan

The Heart and Stroke Foundation of Canada has set out four high-level pillars for a nationwide campaign to improve response and outcomes for out-of-hospital cardiac arrest. Heart and Stroke in Atlantic Canada, which covers the provinces of Nova Scotia, Prince Edward Island, and Newfoundland and Labrador, requires a process that will translate the high-level pillars into a strategic and tactical advocacy plan that is tailored to each province’s current state in terms of legislation, political and public receptiveness, and infrastructure, as well as their anticipated needs over the next five years. The interns will conduct a thorough analysis of the current state within each province with regards to cardiac arrest response and policy, leverage political relationships that can enable systematic change, and create strategic campaigns in line with the specific needs and advocacy goals of each province, creating an evidence-driven and actionable roadmap for Heart & Stroke in Atlantic Canada to successfully carry out the campaigns.

Voir la description complète du projet
Superviseur du corps professoral :

Katie Dainty;Timothy Chan

Étudiant :

Partenaire :

Heart and Stroke Foundation (NS)

Discipline :

Business

Secteur :

Other services (except public administration)

Université :

University of Toronto

Programme :

Business Strategy Internship

Stream Data Analytics Pipeline Based on IoT Big Data Environment for Safer Fleets and Smart Cities

Road safety affects everyone, not just Geotab customers. With several years of driving and environmental data
from over 2 million connected vehicles we have an opportunity to make our customers safer, as well as our
communities and cities. To reduce accidents, we need to understand both the driving behavioural patterns that
are predictive of accidents, and the environmental factors involved. To achieve this, the data infrastructure
should be capable of processing a series of real-time telematics data, as well as time-series historical records
generated from existing machine learning models to respond to real world incidents within a short period of
latency.

Voir la description complète du projet
Superviseur du corps professoral :

Ben Liang

Étudiant :

Partenaire :

Geotab Inc

Discipline :

Computer science

Secteur :

Information and cultural industries; Professional, scientific and technical services; Transportation and warehousing

Université :

University of Toronto

Programme :

Accelerate

Face Matching 1 to N

Fraudsters and money launderers have no place in today’s digital economy. To protect against fraud and financial crime, businesses online need to know and trust that their customers are who they claim to be – and that these customers continue to be trustworthy. Jumio uses the power of AI, biometrics, machine learning and certified liveness detection to help you rapidly convert more customers, stop fraudsters from infiltrating your online ecosystem and get in compliance with KYCIAML. The main goal of this project is to identify an individual from a face image by searching in a gallery of stored face images. The idea is to employ machine learning techniques to solve this problem. For this purpose, we want to review and evaluate existing solutions in order to develop an in-house approach. The fundamental challenges to tackle are: the particular conditions of our data, the huge amount of data to be handled and the quick response time demanded by our use case.

Voir la description complète du projet
Superviseur du corps professoral :

Ioannis Mitliagkas

Étudiant :

Partenaire :

Jumio

Discipline :

Computer science

Secteur :

Professional, scientific and technical services

Université :

Université de Montréal

Programme :

Accelerate

Efficient Multi-frame Obstruction Removal

Modern mobile phones have become the dominant photography device in recent years. However, their users are often not professional photographers. Thus, they lack the skills of choosing proper lighting, proper shot framing, and proper settings on the camera. In particular, photographs are often taken in unfavourable conditions where the scene of interest is obstructed by a fence or a window. We would like to remove such obstruction automatically. Capturing scenes from multiple viewpoints, not only helps to identify the obstruction better, but also helps in removing it. There are existing multi-frame obstruction removal methods, but they are too computationally intensive. This project will focus on a multi-frame obstruction removal approach that is accurate while keeping in mind the computational constraints that would enable mobile deployment. This is of major interest to Samsung as one of the leaders in the mobile phone industry.

Voir la description complète du projet
Superviseur du corps professoral :

Kiriakos Neoklis Kutulakos

Étudiant :

Partenaire :

Samsung Electronics Canada

Discipline :

Computer science

Secteur :

Manufacturing

Université :

University of Toronto

Programme :

Accelerate

Community-based monitoring framework development for the Foxe Basin Kivalliq North Sapujiyiit/Guardians of the Sea Society

The proposed research project will be to provide an understanding and overview of Inuit women’s concerns, feelings, and thoughts regarding how their day to day lives are changing due to climate change and increased development in Nunavut. The partner organization (Sapujiyiit Society) will have the role of guiding the HQP through formulation of the research question, project design and implementation, assistance with research objectives and incorporation of perspectives of Inuit women. This work is aligned with government initiatives outlined in the throne speech and will help us understand the impacts of climate change on women. Outputs will be publicly available to be adopted and modified by other guardians programs.

Voir la description complète du projet
Superviseur du corps professoral :

Brent McKnight

Étudiant :

Partenaire :

Sapujiyiit Society

Discipline :

Sociology

Secteur :

Professional, scientific and technical services

Université :

McMaster University

Programme :

Accelerate

An automatic syringe filling system that uses robotic arms controlled by deep learning algorithms to help pharmacy technicians with their tasks

In certain scenarios, in the hospitals, oral medications need to be administrated in a liquid form using syringes in order to avoid complications such as asphyxiation due to the blocking of airways, which is common in children. Consequently, a pharmacy technician may need to fill up hundreds of syringes per day for different units at the hospital such as pediatrics, geriatrics, and psychiatry.
These sets of repetitive tasks expose pharmacy technicians to high risks of work-related musculoskeletal disorders. For instance, tendinitis may occur during syringe filling due to poor wrist posture. In addition, the shortage of personnel is hitting our hospital pharmacy departments hard and is forcing the hospitals to make painful choices in the services they can offer. To address this problem, we will be using robots capable of integrating Deep Learning (DLs) algorithms. This robot can help pharmacy technicians with their tasks. As a result, pharmacy technicians can be assigned to other duties where their presence is needed.

Voir la description complète du projet
Superviseur du corps professoral :

Usef Faghihi

Étudiant :

Partenaire :

Kobotik Inc.

Discipline :

Computer science

Secteur :

Health and Related Sciences & Technology; Professional, scientific and technical services

Université :

Université du Québec à Trois-Rivières

Programme :

Accelerate

Développement d’une méthode non-supervisée d’apprentissage profond pour la détection de défaillance à partir de signaux acoustiques et vibratoires

L’objectif du projet est d’automatiser la reconnaissance des signaux acoustiques et vibratoires a l’aide d’un modele d’apprentissage profond. Cette reconnaissance est la premiere etape du developpement d’un modele de pronostic et de suivi de !’evolution de la degradation des alternateurs. Ces signaux sont issus de mesures effectuees periodiquement sur les alternateurs durant leur fonctionnement. Les mesures acoustiques etant considerees comme non-intrusives par rapport a !’exploitation des alternateurs, leurs etudes et analyse a l’interieur d’un algorithme de detection automatique de defaut aura un impact majeur relativement aux strategies de priorisation de la maintenance. Entin, les signaux vibratoires et acoustiques permettent de quantifier l’etat de sante des alternateurs, comme les signaux electrocardiogrammes « ECG » pour l’etat de sante d’un humain.

Voir la description complète du projet
Superviseur du corps professoral :

Ioannis Mitliagkas

Étudiant :

Partenaire :

Institut de Recherche Hydro-Québec

Discipline :

Computer science

Secteur :

Professional, scientific and technical services; Utilities

Université :

Université de Montréal

Programme :

Accelerate

Data Drift Detection and Monitoring

When a previously trained machine learning model is put into production, the production phase begins where said model makes predictions on the inputs provided to it. When the distribution of production data changes over time, we talk about data drift. Then the model is likely to become less efficient, or even obsolete. The project consists of building an intelligent system capable of alerting in the event of a data drift that would have a significant impact on the system.

Voir la description complète du projet
Superviseur du corps professoral :

Heng Li;Foutse Khomh;Mohammad Hamdaqa;Maxime Lamothe

Étudiant :

Partenaire :

Zelros;Zelros (France)

Discipline :

Computer science

Secteur :

Professional, scientific and technical services

Université :

Polytechnique Montréal

Programme :

Accelerate

Précédent
1...1,3931,3941,3951,3961,3971,3981,399...2,635
Suivant

Joignez-vous à un écosystème d’innovation florissant.

Abonnez-vous à notre infolettre dès maintenant!

S’inscrire

Mitacs — Retour sur innovation

Mitacs est financé par le gouvernement du Canada, le gouvernement de l’Alberta, le gouvernement de la Colombie-Britannique, Research Manitoba, le gouvernement du Nouveau-Brunswick, le gouvernement de la Nouvelle-Écosse, le gouvernement de l’Ontario, Innovation PEI, le gouvernement du Québec, le gouvernement de la Saskatchewan, le gouvernement de Terre-Neuve-et-Labrador et le gouvernement du Yukon.