Multi-modal learning of human pose representation for conditional motion synthesis

The goal of the project is to generate realistic human movement in 3D animations. This is important to make movement animations in games and movies appear real. Typically, creating high quality animations is a resource and time consuming process that requires the participation of human actors in motion capture sessions. In this work, we present […]

Read More
Improving Monitoring and Decision-making with Uncertain Sensor Data

Terrestrial contaminated sites – such as abandoned oilfields, chemical spill sites, or former industrial zones – are a major environmental problem in Canada and around the world. Environmental Material Science has created new environmental monitoring equipment that generates high-resolution spatially and temporally explicit data on environmental quality. The data must be visualized and then used […]

Read More
Prediction of biological age based on DNA-Methylation patterns

DNA methylation is a biological process that involves the chemical modification of DNA with methyl groups at specific regions of a chromosome. DNA methylation and demethylation change the processes associated with turning on and off genes. Changes in DNA methylation patterns can be influences by environmental forces which are recorded at the molecular level and […]

Read More
ChrysaLabs : Prédiction du rendement agricole

ChrysaLabs développe des solutions technologiques innovantes pour améliorer l’intégration des connaissances sur la santé et la fertilité du sol dans les pratiques agricoles. L’un des défis les plus importants dans l’amélioration des pratiques agricoles est de comprendre l’impact des changements de pratiques sur l’amélioration du rendement et sur l’environnement. L’objectif de ce projet est de […]

Read More
High-throughput linguistic content comparison and sentiment analysis

Scrawlr is a platform for unconstrained, global interaction with all internet content and users. Scrawlr allows for user evaluation and unconstrained classification of any Scrawlr-hosted or non-Scrawlr content. For non-Scrawlr content, this evaluation and classification allowance will be first at the URL level but will subsequently be provided at the individual content component level. Scrawlr […]

Read More
Weakly Supervised Behavioral Modeling for Controllable AI Agents in Video Games

The project aims at developing a new type of Reinforcement Learning algorithm that would allow to retain more control over the artificial agent once its training is completed. This framework would combine modern unsupervised modeling techniques to capture the variability of a set of demonstrations and user-defined programmatic functions that can characterise particularly important factors […]

Read More
Forex data gathering and analysis acceleration

A prototype Forex currency trading system has demonstrated a proof of concept on a static data set. The current project proposes to implement the trading system for real-time applications. Foreign currency trading (Forex) is a high risk activity where low stakes multiply out to large gains and large losses.The trading activity is time dependent and […]

Read More
AI Enabled Subnetwork Selection

Cancers are heterogeneous disease that hijack many of the body’s normal biological processes. Additionally, tens of thousands of genes are involved in each person’s normal biology, while only a fraction of those are repurposed by cancers to drive disease. At an individual level, utilizing entire transcriptomes is rare, as there is too much information for […]

Read More
Pavemetrics : Inspection de voies ferrées

Pavemetrics, chef de file dans les systèmes de vision numérique pour l’inspection des infrastructures routières, ferroviaires et aéroportuaires, a développé de nombreux capteurs couplés à un logiciel de traitement de données 3D afin de les offrir au monde entier. Parmi les clients de Pavemetrics, on retrouve principalement les intégrateurs de services, les firmes de génie […]

Read More
Thales : Expliquer l’impact des graphes de connaissances dans les systèmes VQA

La réponse visuelle aux questions (Visual Question Answering VQA [1]) a été introduite pour combler le fossé entre le traitement du langage naturel et les applications de compréhension des images dans l’espace commun de la vision et du langage. La plupart des benchmarks VQA calculent une représentation de la question en utilisant des techniques d’intégration […]

Read More
Bentley : Détection, classification et inspection de panneaux routiers en contexte semi-supervisé

La quantité et la diversité des panneaux de signalisation routière sont en augmentation constante. Parmi les défis posés par cette situation, on retrouve celui de l’inspection. En effet, ces panneaux doivent conserver des propriétés physiques déterminées (comme une réflectivité minimale) afin de satisfaire aux exigences gouvernementales. L’inspection manuelle de chaque panneau étant fastidieuse, des systèmes […]

Read More