Phishing emails are a common form of a cyber attack. Attackers use leading content to fraud victims, such as fake banking information, forged Google Alert messages, and so on. Phishing attacks have been around for years, causing countless serious consequences, such as financial losses and confidential documents leaks. As of now, there is still no […]

The student will develop a recommendation system for restaurants, proposing food selections to customers. This system will be based on GNN models to predict a customer’s need based on both user and order data. The data include previous purchases, data of dishes and similarity of users collected from online food orders. In the first phase […]

Crowds are everywhere and studying them is important for urban planning, transportation, evacuation, and safety at large public events. Simulation modelling lets us study crowd behaviour by simulating large numbers of individual pedestrians and seeing how they behave as a crowd in different, sometimes dangerous scenarios. Simulation allows us to study situations that would be […]

This project aims to detect 3D brain lesions automatically using machine learning and MRI images. Given an MRI image of a brain, the project will automatically detect any size lesion. The project will establish a baseline, as well as improvement over the baseline for which performance metrics will be validated. The partner organization benefits from […]

I am working on reinforcement learning for finance based on deep mathematical knowledge and the host supervisor is working on financial engineering, reinforcement learning and Markov decision process. We will study deep reinforcement learning for stock market trading and for portfolio management. The goal is to find a practical deep reinforcement agent to manage stock […]

Macros can greatly enhance the capabilities and convenience provided in documents. They also invite adversaries to include malicious code in lure documents, often used as initial access into a user’s environment. This project will extract and analyze macros and determine their indent and potential for malicious code execution. Reducing time to response through malicious code […]

In today’s world, organizations protect themselves and their customer’s data through the implementation of complex cybersecurity solutions composed of many different nodes, each generating constant streams of data. Building reports from this data through the calculation of various metrics can provide much needed visibility into the state of the environment. However, building such reports can […]

Malicious websites in general, and phishing websites in particular, attempt to mimic legitimate websites to trick users into trusting them. The goal of the project is to develop algorithms for detecting these malicious websites in two contexts: • detecting if a site visited by a user is a malicious site • detecting malicious sites that […]

Most collaborations and government departments share their threat data feed in Data Exchange. Inescapably, nowadays with increasing threat data, it is a challenge to extract a large amount of threat data and unify the format more quickly. And as more and more companies join in sharing, the redundancy of this duplicate data will increase dramatically. […]

I will apply the cycle method on story visualization. And I will apply the technique about learning unpaired data, unsupervised learning to improve general performance.

Le stagiaire entreprendra la tâche d’améliorer la robustesse des systèmes de dialogue de la Banque nationale. Afin qu’un système de dialogue soit apprécié par les utilisateurs, il est primordial que cette technologie réponde de façon appropriée à une question. Une erreur propagera de la fausse information et diminuera la crédibilité de l’organisation, deux graves problèmes. […]

We present a method for automatically estimating the lighting conditions from a single image. As opposed to most previous works which proposed methods that estimate only global lighting or use a limited illumination representation (low frequency SSH, parametric model), the proposed method attempt to use a new spatially-varying light representation with realist texture to render […]