Online Business Model Prediction Service

In the business, it is critical to understand and predict needs of customers in advance and in precision. Machine Learning and Artificial Intelligence make it possible to extract the desirable properties and predict the objective in the future. This project is interested in the implementation of this concept as a toolbox. The toolbox will consider Online Business Model Prediction Service (OBMPS) as the objective to create an environment to drive smart decisions.

Study of the Latent Space in NLP: Mathematical Foundation and Application to Disentanglement

Recent progress on word and sentence embeddings has enabled efficient representation and learning of complex high dimensional probability distributions over rich text data. The proposed research aims at addressing some of the fundamental questions in this field: What are the natural mathematical structures on that latent spaces? How to find a meaningful basis? What is the best method of disentanglement for NLP?

Development of machine learning and artificial intelligence toolbox to monitor data center risks and performance

Missing Link Technologies Ltd. provides customized and effective solutions to its partners for their IT infrastructure & operations. They are especially involved with solving the IT facilities management challenges of Telcos and Datacenters. The proposed research in collaboration with Missing Link Technologies Ltd. is aimed at leveraging the advances in machine learning and artificial intelligence in risk management, performance assessment and availability prediction of a data center. A suite of tools and techniques will be proposed towards this end.

Advanced Signal Processing and Machine Learning for BLE-based Indoor Localization

The Internet of Things (IoT) is a new emerging paradigm and is rapidly gaining ground in different applications of significant engineering importance including but not limited to smart buildings, and smart public environments. The main enabling factor of this promising paradigm is integration of identification, localization, and navigation technologies with smart hand-held devices equipped with sensing, processing, and communication capabilities.

NLP Sales Assistant

The goal of this project that will be conducted in collaboration with Heyday is to create a technology that uses a given messaging platform (e.g. Facebook Messenger, web chat widget) that allows users to communicate easily and smoothly with their preferred brands or retailers. This technology should allow the automation of answers and interaction between users and retailers. The technology that we would like to develop will be based on advanced Natural Language Processing (NLP) and machine learning techniques.

An information-theoretic framework for understanding generalization in neural networks

Deep neural network (DNN) is a class of machine learning algorithms which is inspired by biological neural networks. DNNs are themselves general function approximations, which is the reason they can be applied to almost any machine learning problem. Their applications can be found in visual object recognition in computer vision, translating texts in unsupervised learning, etc. DNNs are prone to overfitting because DNNs usually have many more parameters than the available training data. However, they usually have a low error on the test data.

Measuring investments performance using Artificial Intelligence in impact investing and socially responsible investing

In the last few decades, Socially Responsible Investments (SRI) have growingly become a relevant issue. This trend is expected to continue and strengthen going forward. Moreover, the steady trend of rationalization and marketization of the social sector over the past four decades has set expectations and norms of rationalized practices such as Social Performance Measurement (SPM) encompassing practices such as impact evaluation, outcome measurement, and program monitoring adopted by an organization to measure its progress toward its social goals, reflecting internal influences.

Automatic Tracking, Localization and Orientation of Hockey Players, Using Broadcast Videos

Automatic analysis of sport videos is an attractive research area in computer vision that is driving the sport analytics towards a more technological edge. By automatically analyzing sport videos, lots of information could be drawn that benefits the teams, coaches, referees, players and even the fans, such as: extracting strategy of the game, technique and performance of each individual player, performance of the referee in a competition, and etc. This area of research, although attracted many researchers in computer vision community, is still in its infancy.

AI-Based Automation of the Candidate Recruitment and Management System for TAS

TAS (Techno Aero Services Inc.) is a Canadian recruitment agency specialized in aeronautics, engineering, and technology. At TAS, candidates currently submit resumes to a database through a web interface. These resumes are then manually processed by recruiters before a suitable candidate is matched to a position. Through this laborious manual process, great prospective candidates often get lost in the piles.

A new rapid method for clay, silt and sand-sized mineral particle size measurements of soils and soil products

Information on the proportions of clay, silt and sand-sized mineral particles in soils and in soil products used for construction is critical for understanding their physical and chemical properties and for their proper use and management. However, these soil mineral particles are usually glued together with inorganic cements (calcium carbonate, iron and aluminum oxides/hydroxides) and organic matter which can make measurements of the proportions of the individual mineral size separates using current methods problematic.

Pages