Innovative Projects Realized

AI-assisted Recommendation System for False Positive Reduction at Security Operations Centers

Security Operations Centers (SOCs) are responsible for detection and review of malicious interactions. The SOC issues tickets for interactions that are considered suspicious or threatening. These tickets are then inspected by analysts for approval. For sake of safety, this ticketing system often issues too many “false positives”, i.e., it alerts for interactions that are not really threatening. While this keeps the security level high, it can cause analyst fatigue due to high volume of unnecessary ticket reviews. This project aims to develop an AI-assisted system to refine detection mechanisms at SOCs and reduce the issue of unnecessary alerts. This can contribute significantly in enhancing SOC efficiency by both decreasing the number of false positives and reducing the number of reports being processed by the analyst in a certain time period.

Exploring How Mathematics and Literacies Homework is Taken Up the Context of After School/Summer Programs

Our current research is focused on homework policies, how teachers enact such policies, and how homework impacts family life. Extending that work into this proposed MITACS research project, we aim to examine after-school and summer programs and how such programs are connected to schools and homework. This project will provide insights into how mathematics and literacy are taken up outside of the context of school. The partner organization will provide a space for an after school/summer program where an intern will work with children in mathematics and literacy. The project will recruit an intern to help organize and support new and existing programs with the partner organization. The proposed research aims to: a) investigate how mathematics and literacies homework is taken up the context of after school/summer programs (b) examine students’ perceptions of what constitutes math and literacy and (c) support learning during the summer months.

Research intern for leaf morphodynamics

The ability for movement to adjust posture and growth in response to environmental stimuli is important for plants as sessile organisms. This project aims to understand how mechanics regulate plant movement by studying leaf movement in Arabidopsis. This research combines cutting-edge imaging and mechanical modeling techniques from both the host institution, Université de Montréal, and the home institution, Nara Institute of Science and Technology, to explore the driving mechanism of the movement from a multi-scale mechanical perspective. This project strengthens international collaboration between the two institutions and promotes research in multiscale morphodynamics, contributing to advancing the field of plant biomechanics.

Revue systématique sur l’utilisation des technologies en prévention des troubles des conduites alimentaires : portrait des approches actuelles et recommandations

Le projet vise à identifier les modalités technologiques (ex., application mobile, messagerie instantanée, jeux sérieux, etc.) utilisées dans les interventions existantes pour prévenir les TCA; et comparer leurs efficacités. De plus, ce projet examinera les différents publics cibles et les caractéristiques des individus ciblés par les interventions numériques existantes pour prévenir les TCA. Cela inclut, par exemple, l’âge, le sexe et le rôle (ex., parent, éducateur, enfant, etc) des individus ciblés. Nous comparerons également l’efficacité des interventions préventives selon ces différents contextes. Les retombées potentielles du projet incluent la génération des connaissances qui pourraient mener à des répercussion concrètes en TCA, où les besoins en prévention sont incontestables et une meilleure compréhension comment la technologie peut être mise à profit comme outil de prévention en TCA. De plus, nous serons également en mesure d’identifier les lacunes et d’émettre des recommandations quant aux approches optimales, en plus de clarifier si l’efficacité et la pertinence des interventions varient en fonctions des caractéristiques de leurs public cibles. Le stagiaire sera capable d’acquérir et d’améliorer ses compétences en recherche dans des domaines variés, dont la psychiatrie, la méthodologie et la technologie en développant plusiers aptitudes essentiells en recherche.

Conformal prediction, fairness and calibration

The internship focuses on the intersection of mathematics, machine learning, and ethical AI, specifically within the domains of conformal prediction, fairness, and calibration. Conformal prediction is a statistical framework that provides mathematically rigorous confidence measures for machine learning predictions, ensuring that the uncertainty quantification is valid under minimal assumptions. In this project, the goal is to explore how conformal prediction methods can be extended or adapted to meet fairness criteria, addressing biases that may arise in datasets or prediction algorithms.

Analyzing process data alongside traditional item responses to obtain more accurate imputation, offering a deeper understanding of respondent behavior and enhancing the quality of imputing missing responses

This project aims to enhance proficiency estimation in large-scale assessments by improving missing-data imputation techniques. Specifically, the study focuses on refining Multiple Imputation with Denoising Autoencoders (MIDAS)—a deep learning-based approach—by incorporating item response time as an additional contextual feature. Unlike traditional item response theory (IRT) or regression-based methods, which rely on strong assumptions, the proposed approach leverages the flexibility of deep learning to better handle the complexity and dimensionality of large-scale assessment data.

AI/ML in Applied Marine Bioacoustics: Exploring the transfer of existing models from other domains

This project aims to answer the research question “can existing AI/ML models from other domains be applied to help address marine bioacoustics challenges?”

One of the key challenges is that marine bioacoustics lags behind terrestrial bioacoustics in the level of research attention and technical advancement. Additionally, bioacoustics as a field has been slower in leveraging AI/ML techniques compared to other domains, such as speech recognition and medical imaging.

The Mitacs intern will:
1. Expand the preliminary literature review undertaken during Winter 2025 through a Memorial University Professional Skills Development Program 60-hour Global Student Exchange placement.
2. Curate and clean a database of known/identified sound recordings for selected marine species, as well as recordings of marine environments containing many sounds (ambient, shipping, marine mammals, fishes, etc.).
3. Use the cleaned data to test the effectiveness of existing AI/ML models transferred from other domains.
4. Engage experts for challenge identification/confirmation, species selection, data provision and results verification.

newkid x Patricia: Stepping into the front doors as the new kid

newkid® is a creative company specializing in branding for startups that want to stand out. We establish and evolve brands to have a strong sense of self, clarity of purpose, distinctive style, and a singular perspective. Our work spans across brand strategy, creative direction, visual identity, naming, advertising, packaging, and digital experiences.

The challenge: As the demand for brand design services grows, startups expect more innovative, efficient, and dynamic branding solutions to differentiate themselves in highly competitive markets. Traditional branding processes can be resource-intensive, requiring extensive research, exploration, and iteration. Additionally, newkid aims to refine its internal design methodologies and leverage new tools to enhance efficiency while maintaining the quality and creativity we are known for.

Collecte de données et prototypage de modèles d’intelligence artificielle pour la détection d’anomalies acoustiques

Le projet vise à optimiser les performances des raboteuses industrielles utilisées dans la production de bois d’œuvre. Cette initiative répond à des défis croissants, tels que la pénurie de main-d’œuvre qualifiée et les dysfonctionnements fréquents des équipements. Il se concentre sur la collecte de données acoustiques provenant de microphones et de capteurs comme des accéléromètres et des capteurs d’émissions acoustiques. Ces informations permettront de développer des modèles d’apprentissage automatique capables de détecter automatiquement les anomalies, facilitant ainsi l’identification des problèmes mécaniques avant qu’ils n’affectent la qualité ou le rendement. En aidant les opérateurs moins expérimentés à comprendre leur machinerie, ce projet vise à réduire les pannes imprévues. Réalisé en collaboration avec Bois Daaquam Inc., le projet comprend deux étapes principales : la collecte de données multisource sur site, suivie du prototypage de modèles intelligents basés sur ces données. Ces travaux apporteront des bénéfices non seulement à l’entreprise partenaire, mais également à l’industrie canadienne en générant des outils applicables à d’autres scieries.

Application of Generative AI for Business Intelligence and Data Analytics

This project focuses on implementing Generative AI capabilities for WebPal’s data warehouse system at Palomino System Innovations Inc. The initiative aims to enhance business intelligence and data analytics capabilities through an intuitive AI interface that simplifies complex data management tasks.

Business & Operations Growth

Turolight is a leading provider of energy-efficient lighting solutions, specializing in cutting-edge Light Emitting Diode (LED) technologies tailored for commercial and industrial applications. The company offers a comprehensive range of high-performance LED products and smart lighting systems designed to meet the evolving needs of sectors such as manufacturing, warehousing, retail, and institutional facilities. With a strong commitment to sustainability, Turolight aims to enhance lighting performance by improving brightness, color accuracy, and uniformity while minimizing glare and ensuring consistent illumination across various environments. This involves optimizing lighting for specific tasks and integrating adaptive controls that respond to occupancy, daylight availability, and user preferences. At the same time, the company strives to reduce energy consumption and environmental impact through the use of high-efficiency components, intelligent controls, and environmentally conscious materials.

Intelligent modular electromagnetic mapping instrument for non-contact material characterization

The main objectives of the project are designing, developing, and integrating a high-resolution electromagnetic mapping instrument, with interchangeable sensor arrays or sensor suites for single-sided access, non-contact characterization of materials, by evaluating different electromagnetic properties (i.e., electric conductivity, magnetic permeability, dielectric constant) while using an intelligent modular architecture.
Sustainability of new and improved manufacturing methods, one of the main objectives of the Advanced Manufacturing cluster program, can only be assured though repeatable processes and quality of the resulting parts. In the aerospace industry, introduction of new manufacturing processes (i.e., additive manufacturing) and novel materials are required to demonstrate fitness for purpose. Sensing is simultaneously at the heart of characterization of new materials and maintenance of existing structures. Electromagnetic sensing has a wide variety of applications, as electric and magnetic material properties could identify discontinuities that are or would be detrimental to a part, component, or structure under certain operating conditions or in certain environments. the goal of this project is to develop and integrate electromagnetic sensing instrumentation that would allow for fast, accurate, and non-destructive evaluation of metallic and non-metallic structures, with multi-technique and multi-sensor competencies, that could be deployed both in the lab and in an industrial setting.