Projets novateurs réalisés

L2M – Reinforcement-Learning-Driven Electronic Design Automation (EDA) for Optimal Layout Placement

Reinforcement Learning (RL) driven Electronic Design Automation (EDA) is revolutionizing layout placement optimization for integrated circuits, enabling a faster design process. By incorporating the RL techniques, we enhance the historically precise yet labor-intensive process for smart integrated circuit (IC) fabrication. This innovative approach streamlines IC design, accounting for factors such as foggy and proximity effects, promoting both efficiency and accuracy in the layout placement design.

Optical diagnostics for probabilistic quantification of defects in functionalized 2D nanomaterials

This project aims to quantify defects in the probability framework for 2-dimensional (2D) materials with superior performance using optical techniques, primarily focusing on molybdenum disulfide (MoS2). Additionally, defects are an important factor influencing the performance of single-photon emitters (SPEs), a type of quantum device. This project will directly relate the SPE performance with defect concentration. Thus, the second part of the project involves fabricating SPEs based on Argon-ion irradiated/ strain-induced defective MoS2 and correlating device performance with a predeveloped statistical inference prediction model. This project will benefit participating institutions by advancing the development of 2D material applications in the quantum field and enhancing the understanding of material defects. This research project will help to build a foundation for large-scale production in the future.

Development of a robust, low-complexity infant cry classification system

This proposal aims to develop a robust, low complexity infant cry deep learning classification based on various babies’ responses to physiological needs such as hunger or to discomfort and pain. The significance of this research lies in its potential to enhance early detection of needs and moods in newborns, contributing to improved infant care, early intervention and augmented infant-parent communication.
The novelty of the proposed research lies in applying methods to improve performance when small datasets are available and to reduce complexity in deep learning classification systems for infant cries. In this project we will also select and benchmark multiple datasets for training and testing, evaluate and compare different methodologies for feature selection and scaling, and implement a model suitable for real-time applications.
Future research will be extended in subsequent years to include the detection of additional emotional responses, and include babies who are diagnosed with medical conditions, but the emphasis of this one-year research proposal is on machine learning, engineering, and computer science aspects, and it will be performed using existing public domain datasets.

L2M – Satellite Monitoring, Analysis, and Reporting Tool for Harmful Algae Bloom identification: introducing SMART-HAB, a machine-learning tool to identify and visualize harmful algae blooms in near-real time.

Harmful algal blooms (HABs) are a growing threat to drinking water, fisheries, public health, and recreation. In recent years, HABs have increased in frequency and severity in both freshwater and marine environments. Blooms are hard to monitor because they can occur unexpectedly, and reporting methods across Canada are inconsistent, creating a patchwork of alerting methods for industry and public sectors to rely on. Funding to monitor HAB activity is increasing, but current methods are expensive and time consuming. Field teams are limited by water quality testing capacity, and even buoys with remote sensors monitoring water quality can only test water that interacts with the device.

Our team is developing an addition to the HAB monitoring and alert network with the Satellite Monitoring and Reporting Tool for HAB identification (SMART-HAB). SMART-HAB uses satellite imagery with high temporal and spatial resolution to identify potential HABs in Canadian waters using machine learning algorithms. This application can greatly increase monitoring groups’ capabilities by showing users where HABs are occurring, saving both time and resources. SMART-HAB has experienced success in estimating bloom severity, and has been positively correlated to in-situ measurements of cyanobacteria blooms. We hope that SMART-HAB can be used across the country to create a consistent, widespread monitoring network, and alert users in near-real time of HAB activity, severity and extent. With the help of Mitacs and Springboard Atlantic Inc., we are refining our product for future users to prepare SMART-HAB for market.

L2M – Enabling knowledge transfer between science education and coastal communities by leveraging generative AI and climate science publications

There are over 250 million scientific publications and reports with an increasing rate published each year, yet many are not accessible to the public due to their technical language and content hidden behind paywalls. This project aims to leverage AI (Artificial Intelligence) and a curated database of ocean-climate literature to enable educators and students to think critically and ask challenging questions about how climate change will affect their communities, as well as empower them to engage with and apply scientific knowledge towards climate solutions. But before we start designing a system, we need a real-world understanding of who educators are and what barriers exist with finding evidence. With secondary and post-secondary educators in Atlantic Canada as an initial population, a study focusing on the context of actors and the socio-cultural-organizations is needed to understand information needs, work demands, perceptions, and how they currently search for information.

Fast Scenario Identification and Classification

Self-driving cars represent a transformative innovation in transportation, promising safer and more efficient travel. However, their development faces significant challenges, including accurate prediction, path planning, and safe maneuver execution, especially under varying driving conditions. Ensuring safety across all potential scenarios within the operational design domain is paramount. To effectively address this, we propose developing algorithms that are both cost-effective and low-latency. This would enable faster processing and identification of specific scenarios that are rare or of particular interest to autonomous vehicle developers.

Optimisation et contrôle des flux d’air dans les bâtiments du CIMEC

In recent years, factors like thermal comfort, indoor air quality, and energy consumption in school buildings have received increasing attention. Lecture theatres and chemistry labs pose significant challenges for HVAC design due to thermal stratification in lecture theatres and ineffective ventilation in chemistry labs, caused by large spaces and pollutant generation, respectively. Despite many studies on these topics, there’s a lack of research on thermal stratification in cold climates like Quebec, where the issue is severe in winter, and on the effectiveness of local ventilation methods in chemistry labs.
To address these challenges, this project will conduct field tests and numerical simulations to analyze air distribution in an existing lecture theatre and a planned new chemistry lab in a Quebec college. The goal is to optimize indoor environment conditions and energy efficiency for both spaces. The outcomes will enhance thermal comfort, energy efficiency, and air quality, providing guidance for the design and control of air distribution, aiding renovations and new developments. This will help schools in Quebec and across Canada achieve healthier indoor environments, better thermal comfort, and greater energy efficiency.

L2M QC 2024 – “Brave Technologies : A new breast cancer diagnostic device”

Breast cancer remains a major global health concern, with high prevalence rates in new case diagnoses across the world. The primary challenge in breast cancer detection centers on achieving early and precise diagnosis as actual methods have limitations. The primary challenge in breast cancer detection lies in achieving early and accurate diagnosis. Current methods face significant limitations, including difficulties in accommodating the variability in tumor characteristics such as size, shape, and location. Our system offers substantial opportunities to address these issues effectively by providing a non-invasive, patient-friendly, and precise detection method, particularly advantageous for women with dense breast tissue. By introducing our device to the market and building trust with customers, we will be able to detect breast cancer at early stages, providing patients with a better chance to fight and overcome this disease, thereby contributing to reducing mortality rates caused by breast cancer. The main objective of this project is for the intern to conduct an in-depth market study in the Final Users, private sector, and non-profit entities segments for the potential applications of the BRAVE System, an innovative, accessible, and affordable breast cancer detection method specifically designed for women with dense breast tissue

Lossless Ethernet

Telecom operators are demanding more flexible, scalable and energy efficient products. As a system integrator managing networks for operators, Ericsson realize that the benefits of using modern and improved systems is becoming even more critical, as they offer a reduced Operational and Capital Expenditure (i.e. OPEX and CAPEX) of Information and Communication Technology (ICT) systems. The goal of this project is to develop a mechanism to prevent congestion in the network, prevent packets from being dropped and to guarantee delivery of packets between collaborating end-hosts. For such mechanism to be used in a high traffic network, a number of characteristics must be evaluated against a number of constraints and in different network deployment scenarios

Visual Regression Testing of Video Games Using Foundation Models

During visual regression testing of video games, automated techniques are employed to identify visual bugs. Visual bugs can be detected by just looking at them, such as those related to texture and lighting, but also more complex ones like those related to the physics engine (such as a flying horse) that require common-sense reasoning. The impact of visual bugs varies from minor irritations to rendering a game unplayable. Although human experts can often easily spot these bugs, automated detection is challenging. As a result, visual regression testing usually requires significant resources for manual testing.

This project proposes leveraging foundation models to address the visual regression testing problem. Foundation models are large-scale machine learning models that are pre-trained on very large amounts of data from different domains. This project investigates how foundation models can be leveraged and fine-tuned for visual regression testing of games that were published by Electronic Arts (EA), a world-leading game publisher with over 8,000 game makers. Automated techniques for detecting visual bugs in video games would help reduce the cost of testing for EA. The project’s goal is not to replace testers but to help them find more bugs in less time.

L2M-Enhancing Fertility Treatment: A Mobile Application Utilizing Behavioral Change Strategies

A fertility app is proposed to help couples assess their risk of infertility, create personalized behavior change programs, and track fertility-related biomarkers. Personal, health, and lifestyle information will be collected and synced with smart devices to monitor menstrual cycles, ovulation, stress levels, heart rate, sleep patterns, and physical activity. Data will be continuously analyzed by AI algorithms to predict how lifestyle changes can improve fertility markers. Advanced features will include the integration of blood test results and tracking of basal body temperature and cervical mucus to enhance predictive accuracy. Personalized diet, stress management, and exercise plans will be offered, incorporating exercise psychology principles to improve adherence and fertility outcomes. Additionally, educational content and community support will be provided, empowering users with knowledge and fostering a supportive community. Healthcare appointments can be tracked, and prompts for questions to ask healthcare providers will be given. The partner organization will benefit from this comprehensive tool by supporting individuals in their fertility journey, improving reproductive health outcomes, and gaining valuable insights through data collection and analysis.

Bourse postdoctrale RSMC-Mitacs-iA en santé mentale au travail automne 2024

Le travail agricole comporte certaines contraintes contextuelles relatives, entre autres, à l’économie, aux conditions climatiques et à l’environnement. Ces contraintes de travail engendrent un stress intense chez les propriétaires d’entreprises agricoles. En effet, selon une étude canadienne publiée en 2016, près de 60 % d’entre eux était à risque de vivre de la détresse psychologique. Cette étude n’a toutefois interrogé que 10 agriculteurs québécois (moins de 1% des participants). L’étude la plus récente portant sur la santé mentale des agriculteurs québécois remonte à près de 20 ans et montrait que 49,5 % des participants avaient un niveau de détresse psychologique élevé. Une recherche quantitative sera donc menée pour mettre à jour ces connaissances puisqu’elles sont nécessaires pour mettre en place des interventions psychosociales appropriées aux besoins de ce groupe de population. Cette recherche s’intéresse ainsi au risque de vivre de la détresse psychologique chez les propriétaires québécois d’entreprises agricoles, à leur attitude envers le recours aux services d’aide en santé mentale, ainsi qu’à leur capacité de résilience. Des analyses comparatives seront effectuées entre divers sous-groupes établis selon les caractéristiques des personnes participantes (p. ex. selon leur âge, leur sexe, le type d’entreprise agricole). Les résultats de cette recherche pourraient ainsi contribuer à bonifier les interventions psychosociales à l’égard de ce groupe de population.