Characterizing and Improving the Robustness of Convolutional Neural Networks

Convolutional neural networks (CNNs) are expressive function approximators that play an important role in solving modern computer vision tasks, such as object recognition, and even summarizing images in natural language. Given their broad utility, CNNs have already been deployed in performance-critical systems, such as autonomous vehicles. Unfortunately, these models are vulnerable to subtle perturbations of […]

Read More
Bringing Personalized Recommendation to the Legal Domain

Machine Learning has just started to be applied to the Legal Domain. ROSS Intelligence makes it possible for legal professionals to work faster and more effectively. Advanced Recommender Systems have not been previously applied in the Legal Domain. Yet state of the art models such as ones using Deep Collaborative Filtering have proven to be […]

Read More
Ice Hazard Drift Model Study III

Massive drifting icebergs frequently threaten offshore operations on the Grand Banks because of their massive size, and great mechanical strength. These ice hazards move erratically which complicates efforts to modify their trajectory or undertake evasive action. This MITACS project aims to improve security of offshore workers and help protect wildlife and the environment by allowing […]

Read More
P4 SDN testbed integration

The research will consist of exploring a new language as well as a new paradigm shift in the orchestration and analytics involved in operating a Fiber optical infrastructure equipped with IP routers and Computers. These computers will be equipped with programmable devices that will allow further instructions and detailing about the next generation of internet’s […]

Read More
Characterizing the role of probiotics in physiologically relevant ex vivo and in vivo models of infectious colitis

In the intestines of people living with inflammatory bowel diseases (IBDs), the balance of beneficial bacteria is shifted. Instead, the intestine is overloaded with potentially harmful bacteria – a phenomenon known as dysbiosis. This shift in bacterial populations is believed to be among the key contributors to the onset of inflammation observed in IBD patients. […]

Read More
Regulating Abnormal Connectivity in Posttraumatic Stress Disorder via Real-time fMRI Neurofeedback – Year two

Patients with posttraumatic stress disorder (PTSD) are characterized by decreased prefrontal cortex (PFC) regulation on hyperactive emotion generation regions, such as the amygdala. Real-time (rt)-fMRI neurofeedback allows for localized brain regions to be self-regulated through neuroimaging signal feedback. Recently within our lab, learning to decrease amygdala activation via neurofeedback was shown to normalize the neural […]

Read More
Anomaly Detection in transactions volumes

The objective of the project is to investigate how machine learning techniques can be used to detect anomalies in volumes of transactions. This requires the student to conduct a literature review about the topic as well as experimenting with a subset of selected machine learning techniques. The results from the research could help the partner […]

Read More
Rapid Operations Planning for Space Robotics Using Machine Learning

Few things in space flight are routine. Before each time MDA operates the International Space Station’s famous Canadarm2, thousands of simulations must be performed to ensure the success and safety of the operation. This research intends to streamline the process of operations planning for Canadarm2 by using machine learning to predict key outputs from these […]

Read More
A reinforcement learning approach to establishing a Q&A symptom checker to evolve the performance of the visual diagnosis system for dermatological diseases

The partner is creating artificial intelligence which can help diagnose over 1,300 skin conditions with dermatologist-level accuracy. They are focused on building the functionality that is to be deployed through their app and web-interface that makes it possible to snap a photo, ask questions, and get an instant diagnosis. The partner is very focused on […]

Read More
Ultra-reliable and low-latency communication for industrial use case

Ultra-reliable and low latency communication is increasingly an important aspect of future wireless communications. Specifically, in the context of mission critical communications for large-scale networks of sensors and actuators in automated and/or remote-control applications, low-latency wireless communication with high level of determinism is a vital element. The key performance indicators for such use case are […]

Read More
The Greater Toronto Area Watershed Readiness Pilot Project: Gaging how climate and land use change might impact aquatic ecosystems in urban environments.

Cities are often located near sources of water such as lakes, rivers, and oceans. Due to their location, the effects of climate change and urbanization will introduce unique challenges to aquatic habitats located within cities. These challenges could range from flooding, extreme wind, ice storm damage, extreme heat events, and drought as well as long […]

Read More
Aircraft Modification and Configuration Tracking

Tracking the component configuration and modifications to aircraft within the commercial airline business presents a challenge for manufacturers such as Bombardier with currently available methods. This research problem is significant to the aerospace industry to construct efficient maintenance schedules for different aircraft and to properly evaluate system reliability for safety purposes. The objective of this […]

Read More