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.
The ultimate objective of this research project is to use a form of artificial intelligence to be able to classify and identify images of microscopic particles. Machine Learning is the term applied to this type of process, in which an algorithm is created by the computer software itself (i.e. mostly hidden from human intervention) to complete the task.
Transportation of oil and gas through pipeline networks remain a crucial infrastructure for sustainable economic growth in Canada. Pipeline wear and damage will remain a major concern as it can lead to catastrophic failures causing environmental and economic damage if undetected. For easier detection of damage on a large network of pipelines, an array of wireless radio frequency identification tags was developed for steel pipes. However, the material used for the tags were not suitable for pipes made with polymer composites as the stiffness of the copper could damage it.
In this research we will identify current types of customer, taking into account people who prefer to use a variety of platforms and different preferences in terms of how actively they manage their money. . We will carry out focus groups and interpret the results of a survey in terms of their implications for a set of factors that differentiate between banking customers. Using the factor scores obtained in a survey we will segment into meaningful groups (personas).
In Canada, the transportation sector is the second largest greenhouse gas (GHG) emitter and a large contributor to air pollution emissions, which can cause significant health impacts. Since electric vehicle (EV) does not generate any exhaust emissions, introducing EVs can bring health and climate co-benefits to society. From a life cycle perspective, this study will evaluate the environmental, health and economic impacts of introducing EVs in the Greater Toronto and Hamilton Area.
This project will assess the game demands of wheelchair court sports. While this has been attempted in the past, new methods using inertial measurement units (IMU) allow each push to be identified and offer new ways to analyze these game demands and connect them to key performance metrics. With the help of Own the Podium, and Canadian Sport Institute Pacific, these key performance indicators will become an important for developing and developing elite athletes in wheelchair court sports.
Characterization of the energy distribution of ions generated by the plasma in an Inductively Coupled Plasma Mass Spectrometer (ICP-MS) instrument is necessary for a new ion source in that it influences the ion sampling process, transmission efficiency, focusing, and mass analysis in ICP-MS. These energy distribution phenomena are also analogous to the ion beam that has been generated from an electron impact ionization (EI) source. Similarly, better understanding of the ion beam profile results in a better optimization of the EI source for superior performance.
Moulded pulp, is a packaging material, made from recycled papers. It is used for protective packaging such as egg packaging, fruit trays and coffee cup carriers. For many applications moulded pulp is less expensive and environmentally friendly than plastics and styrofoams, however, due to their high water absorption and low strength, these products are limited to only few packaging products. Biobinder, a biobased binder, has been developed from University of Toronto to imparts water repellency and improves the strength of moulded pulp products.
To design effective and patient-specific cancer therapy, sensitive detection of relapse and distant metastases by non-invasive medical imaging is essential, for which MRI offers tremendous potential due to wide availability of the equipment in clinic and avoidance of ionizing radiation. Although gadolinium-based contrast agents are the most frequently used for MRI, they are associated with nephrogenic systemic fibrosis and brain deposition. Thus, less toxic manganese ions (Mn2+) are exploited as an alternative for tumor detection using MRI.
This project allows a PDF to take on a feature engineer role in which she will work closely with a biomedical engineer to identify novel features based on Tracerys proprietary imaging method for Age-related Macular Degeneration (AMD). The PDF will serve as a vital link between clinical disease and computing, interacting with both academic and industry partners for automated feature extraction and ultimately machine learning and artificial intelligence.