A Machine Learning Framework for Exploring Mortality in Developing Countries with Verbal Autopsies

This research project, backed by Unity Health Toronto and the Centre for Global Health Research (CGHR), aims to explore the use of machine learning in predicting causes of death using verbal autopsy data from low-to-middle-income countries. Verbal autopsy is a cost-effective and efficient method for documenting deaths in regions with limited resources.

Optimizing Cybersecurity Data Platform Operational Cost

We focus on building the systems that allow for the storage and analysis of big data for preventing cybersecurity attacks. This telemetry data is collected from customers by leveraging endpoint, network, and cloud data sources both inside and outside our customers’ networks. Our team then builds products that allows our customers as well as our own employees to make business intelligence queries against this large set of data, to better understand their security posture. As our customers grow, the services which host and serve data will need to scale as well.

Identifying optimal growth conditions for cell-cultured fat

This research project will investigate the optimal environment for creating cultivated fat (real animal fat grown in a
lab, without harming any animals). Cultivated fat will be a key ingredient to help alternative protein sources achieve
the taste, texture, and aroma, and nutritional value of conventional meat; however, cultivated meat/fat companies
face challenges in driving down production prices.

Evaluation of Therapeutic Natural Killer cells

Cellular immunotherapy has yielded numerous clinical benefits in treating cancer. While many successes involve CAR T cells, unfortunately this potent anticancer cell therapy can cause serious and even fatal complications. Another immune cell, the natural killer cell (NK), may lead to effective treatment without unwanted side-effects. Furthermore, the use of stem cell-derived NKs provides additional advantages, including the ease of gene editing and ready availability of an off-the-shelf therapeutic.

Improve the reliability and performance of Vision based Machine Learning models to provide more valuable insights to researchers.

Biomedical research, particularly preclinical research, is a complex and challenging field with a high failure rate of 98% in pharmaceutical research investment. Extracting relevant information from preclinical research papers involves synthesizing information from various sources, which is a demanding task that requires domain-specific knowledge. Natural language processing, specifically Large Language Models (LLMs), has demonstrated tremendous potential in extracting information from unstructured text.

A complementary paper microfluidics and sample preservation approach for on-site and in-lab assessment of semen quality

We propose to the develop a process whereby a semen sample can be initially assessed for sperm motility on-site
followed by preservation of the remaining sample for transfer to a lab for quantitative sperm quality analyses. A
microfluidic device using paper will be developed for the on-site motility test, whereby the device shows a colorimetric
signal visible to the eye correlating to the percentage of motile sperm from a few drops of semen.

An Interactive Dashboard for Human-AI Detection of Anomalous Employee Accounts at Risk of Data Exfiltration

Confidential data is one of the most precious assets large organizations can have and data theft can be embarrassing and costly. In this research we will carry out innovative research on detecting employee accounts that are exhibiting risky behaviour that may lead to leakage of data, whether carelessly or through malicious intent. It is difficult to protect against careless actions of employees, or malicious people masquerading as employees. Machine Learning (ML) models have been used to make identification of anomalous data transfers more efficient.

Cortical changes measured with TMS and MEG following 3-weeks of hand movement training using a novel passive device

Stroke is the leading cause of adult disability, with long-term hand function impairments being the most common symptom among stroke survivors. As hand function plays a vital role in performing everyday activities, individuals with stroke face a significant impact in their quality of life including their ability to maintain independence, employment, and socially connectedness. To improve hand function and overcome challenges from this disability, IRegained® has developed the MyHandTM System, a therapy device that promotes various types of hand movements in a gamified environment.

Modern Plug-and-Play Image Priors for RAW Image Enhancement

Modern mobile phones have become the dominant photography device in recent years. However, due to form factor constraints both the camera sensor and lens have to remain compact, which in turn has a negative effect on the resulting image quality such as decreased resolution and noise. Image restoration attempts to recover a clean latent image from degraded input image(s) in a computational manner. To yield a natural image, one must impose additional constraints on the resulting optimization problem.

Development of novel chemical probes and SARS-CoV-2 antiviral therapeutics through engineered HAI-2 protein inhibitors of TMPRSS proteases

Respiratory viruses exploit the biological activity of human proteins present at the surface of airway cells to allow viruses to penetrate that physical barrier and subsequently replicate in infected cells. It has been established that a family of human proteins, the type 2 transmembrane serine proteases, are directly responsible for this viral entry, but they have been challenging to develop drugs against because they have been unable to be studied in isolation.