Innovative Projects Realized

Turn-Key Robot Motion Planning Solutions

To increase the speed at which companies can develop robots we plan to create a package containing a number of key path-finding programs which can be easily connected with any robot. This will make it easier to produce both commercial and experimental robots without the need to create a path-finding program unique to the robot. We also plan to evaluate the possibility of offering expert assistance in choosing and tuning the available programs included in the package.

Detection of Human Presence/Activity through Radio Frequency Signals with Artificial Intelligence

The goal of this porject is to develop a prototype system for human presence/activity detection through radio frequency signals. There have een some recent promising results reported in the literature regarding such detections through WiFi signals using artificial intelligence-based approaches. The postdoc will focus on reproducing earlier results, then move on to enhance the system to detect some human activities of interest. The partner company would like to design, build and commercialize a line of products based on the developed prototype. Among other applications, one potential application for this product will be the detection of a human falling on the floor which is a useful and relevant service/product in the HealthTech domain.

Evaluation of best practice police resources to assist in the rapid response of missing persons with dementia

Three out of five Canadians with dementia wander, raising concern as to how it can be managed. Strategies, such as GPS, offer options for finding missing persons with dementia and may be a preferred strategy by police. As part of the Finding Your Way® Program with the Alzheimer Society on Ontario, a series of education resources were developed in 2018 to assist in the location of this population by police services. The impact of these resources has yet to be evaluated by participating police services. The objective of this project will be to evaluate the impact of the resources developed through the Finding Your Way® Program to assist in the location of missing persons with dementia among police services in Ontario. It will involve a series of surveys and interviews with participating police services The Alzheimer Society of Ontario wants to enhance the reputation of the Finding Your Way® program, and reduce the risk of people with dementia going missing. Partnering with researchers, such as the intern, to have credibility and have police partnerships will help make that happen.

Improvement and validation of the integrity of a novel commercial software for utility pole design

The conventional utility pole design methodologies used a decade or so ago produced poles that would be considered “safe” but in most cases they were not cost-effective solutions. This is because the resulting poles were usually over-designed, mainly due to several simplifying assumptions and incorporation of various rules-of-thumbs in the design procedures. Such design practices were, however, challenged by various regulators and legal/public agencies. In response to the contested situation, in 1994, the industrial partner (Sonideft Inc.) developed a robust and user-friendly design software based on solid engineering fundamentals, commercially known as “Quick Pole”. However, there are two immediate sets of efforts required to complement this effective software: (i) further scrutiny and testing of the options and results produced by the software by comparing its results against those obtained through two different commercially used finite element (FE) software; and (ii) addition of a new module to the software, so it could design poles using today’s lightweight and environmentally resistive fiber-reinforced composite materials.

Biodegradable Polyurethane Elastomers Comprised of Bio-based Plasticizers

Polyurethane elastomers for consumer applications, are developed with primarily bio-based polyester-polyols, isocyanates and bio-based plasticizers which are specifically tailored to result with the required mechanical, flexibility, and biodegradability characteristics useful for consumer applications, especially for footwear and foam products for furniture and automotive industry.

Establishing the Enabling Conditions Required to Facilitate Indigenous-led Nature Based Greenhouse Gas Offset Projects in Canada

Canada is a signatory to global commitments to reduce greenhouse gas levels and one important way this will happen is through the implementation of nature-based solutions for climate change. One example of natural climate solutions is GHG offsets based on restoration, protection and establishment of forests, wetlands, grasslands and peatland areas. Indigenous Peoples are critical to the success of natural climate solutions. Understanding the conditions under which these projects can be enabled is a critical component of establishing Indigenous-led nature based GHG offset projects at a scale and speed that is meaningful for climate change action.
The proposed research will expand on the concept of enabling conditions by creating a framework of enabling conditions specific to Indigenous-led nature based GHG offsets then testing this framework in various case studies which represent actualized, viable projects.

Mentorship Needs in the Ontario Manufacturing Industry

This project will identify the necessary components for a high-quality mentorship program for workers in the Ontario manufacturing industry. Through review of existing mentorship programs, interviews with mentors and mentees, and a job analysis procedure for the mentor role, this research will assess the mentoring needs of manufacturing workers. The project will produce a task-based training guide for mentors so that they can most effectively engage with mentees in order to maximize recruitment and retention. The project will focus specifically on creating a mentorship program to best support underrepresented groups in the industry such as women, young workers and members of racialized groups. This work will contribute to Trillium core program of work that
manufacturing workforce.

Dynamic Deep Generative Graph Models for Financial Forecasting

Borealis AI has access to a huge amount of financial data related to the stock market and is interested in leveraging recent developments in machine learning to better understand this data. Some potential questions emerging from this data are: (1) Given the closing price of a stock in the recent months, can we predict the stock returns within the next month? (2) If a stock crisis occurs, can we predict and control the spread of the crisis? (3) Given the current stock’s history, can we help reduce the risk of investment?. To answer such questions, we propose a network of related stocks based on their correlated returns. Multivariate statistical models that use tabular representations of time series can capture basic correlative structure, but we believe that innovations in machine learning known as Graph Neural Networks (GNNs) will be able to exploit this with more explicit network representation. We propose to develop a novel GNN-based algorithm on network-structured data to efficiently capture its complex structure.

Plausible Futures: What economic and labour market trends might the City see over the next 3-5 years?

The project aims to identify the plausible scenarios or outcomes of the Covid-19 pandemic associated with the City of Toronto. It will provide insights into how the pandemic is to impact the labour market and organizational operations in various industrial sectors. These insights will support ongoing efforts to revive the local economy and in developing and planning efforts to deal with what the future impact on the City operations.

Enhancing interpretability of gaze-tracking convolutional neural networks

Innodem Neurosciences is developing a visible light gaze-tracking algorithms that can be sued to predic a user’s gaze position on the screen of a mobile device without the need for any third-party hardware. This algorithm leverages various image processing techniques, and relies on the use of convolutional neural networks and computer vision. Enhancing the quality of this gaze prediction network will be the primary goal of the resident scientist over the course of this project. As such, the student will support Innoderm’s AI team in the interpretation and modification of our convolutional networks, help the team better tune the model’s parameters, and ultimately test and analyse the efficacy of the changes to our models.

Food Catering Ontology to Enhance UEAT’s Recommendation System

Integration of an ontology for the representation of restaurants, menus and dishes for the catering industry and construction of a recommendation model using this ontology.

Semi-supervised and unsupervised method to increased database labels in the case of classes imbalances

The project aims to improve the amount of labelled samples in a semi-automatic or automatic manner using AI to impove a CNN performance. We will test various state-of-the-art AI methods, in the context of forest inventory, and select the most effective ones.

The benefits will be significant because labelling is an important but tedious task, in many cases, when working with natural forests, some tree species will not occur as often as others (hence creating a shortage in some classes), also there can be co-species to many other species and they are difficult to identify clearly.