L2M- Potatoleaf Doctor( AI model for potato leaf detection)

Agriculture supports food security for 60% of the global population, with potatoes, the 4th most consumed crop, feeding around 1.5 billion people daily (Afakh et al. and LeCun et al.). Canada produces about 5.7 million tonnes of potatoes each year, making it the 12th-largest producer in the world. The country’s potato exports generate around $1.6 […]

Read More
Understanding and improving compositional generalization in vision-language models

Compositionality — the ability to understand new combinations of familiar parts — is a core aspect of human cognition and reasoning, but current vision-language models such as CLIP still struggle with it. These limitations reduce the reliability of systems that must correctly interpret objects, attributes, and relations in real-world scenarios. Although many methods have been […]

Read More
Amélioration des services publics de la Ville de Trois-Rivières

La Ville de Trois-Rivières, 9e plus populeuse du Québec avec 148?469 habitants, compte sur une équipe de 1?250 employés pour offrir des services publics de qualité à ses citoyens. La Direction des technologies de l’information (TI) joue un rôle stratégique en accompagnant l’administration municipale dans l’identification de ses besoins et l’atteinte de ses objectifs grâce […]

Read More
Mining and Modeling Review-Priority Labels in Gerrit for Predictive Review Analytics

Code review is a critical step in modern software development, ensuring quality, preventing defects, and coordinating collaborative work. Large ecosystems such as OpenStack, which is hosted on Gerrit, receive a high volume of code changes every day, making it challenging for reviewers to decide which contributions should be handled first. In response to this, OpenStack […]

Read More
AgriSim

This project will create a realistic virtual greenhouse where robots can practice picking and pruning. It improves plant physics (bending, fruit detachment), models cameras and depth sensors to produce training data, generates many varied greenhouse scenes, and links simulated robots to Kinova’s control software for end-to-end testing. The result: faster, cheaper R&D, fewer risky field […]

Read More
Vers un avatar pour l’interprétation en langue parlée complétée : innovation et accessibilité pour les personnes sourdes et malentendantes

1. Le Service d’interprétation visuelle et tactile (SIVET) offre un service d’interprétation en milieu sociocommunautaire : langue des signes québécoise (LSQ), pidgin, oral support de signes, LSQ tactile (pour les personnes sourdes-aveugles). D’ailleurs, le SIVET est le plus important des quatre services régionaux d’interprétation au Québec et, malgré ses 80 interprètes, refuse en moyenne 120 […]

Read More
Prompt optimization method to maximize the accuracy and reliability of LLM outputs

Draft & Goal (D&G) is a Canadian artificial intelligence (AI) company developing a no-code platform that automates marketing workflows using AI agents for writing, search engine optimization, and fact-checking. The integration of large language models (LLMs) is central to their software solution: it allows non-technical users to harness advanced AI for content creation. However, unlike […]

Read More
3D Gaussian Splatting Multi-agent simultaneous localization and mapping

This project focuses on three tasks: 1. Develop a compact descriptor for 3D Gaussian splats to enable robust registration between partial sub-maps. Inspired by prior learned descriptors, the method will encode local Gaussian parameters for fast matching under noise and partial overlap. 2. Extend existing multi-agent Gaussian SLAM system into a decentralized framework where each […]

Read More
L2M Lensx AI wearable glass for the visually impaired

LensX: Empowering Independence for the Blind through AI Wearable Smart Glasses LensX is an AI-powered assistive technology designed to support blind and low-vision individuals in navigating the world more independently. The project uses a combination of real-time computer vision and voice-based AI to help users identify their surroundings, read text, recognize objects, and receive auditory […]

Read More
Développement d’une plateforme interactive de diagnostic durable basée sur des agents conversationnels guidés par des modèles causaux et des LLMs.

Ce projet vise à aider les petites et moyennes entreprises (PME) à mieux comprendre et améliorer leurs pratiques de développement durable. Pour cela, nous allons créer une application web interactive qui permet aux entreprises de répondre oralement à un questionnaire. Grâce à l’intelligence artificielle, les réponses seront automatiquement analysées et résumées. L’application pourra ensuite proposer […]

Read More
ContactNB/UNB AI Efficiency Project for the NB Contact Centre Industry

The McKenna Institute and ContactNB are collaborating on an initiative to promote AI adoption and responsible use within New Brunswick’s contact centre industry, aiming to enhance productivity, competitiveness, and sustainability by supporting SMEs with AI integration, workforce training, academic collaboration, and attracting investment in the business services sector.

Read More