The collaboration between Amii and ZeroSound aims to address an invisible problem: noise. An estimated 11 million Canadians work in noisy environments, and workplace-related noise has been linked to hearing loss, high blood pressure, sleep issues, and other health concerns.

This collaboration aims to develop and test reinforcement learning (RL) methods that integrate with ZeroSound’s innovative active noise cancellation technology. ZeroSounds’ noise-suppression panels are designed to counter the wavelengths of a loud environment: not just blocking the noise, but actively eliminating it.

Using Amii’s expertise, reinforcement learning allows the panels to react to unpredictable changes: if a panel is moved or a new noise source is introduced, ZeroSound’s technology can automatically adjust. This RL approach has been promising, achieving noise suppression of approximately 12-15 dB — which could reduce a loud construction site closer to the level of a vacuum cleaner. ZeroSound’s technology could not only make industrial workplaces safer, but also has applications in reducing the sounds of noisy intersections, health facilities, and more.

Reinforcement learning is making water treatment more effective, resulting in lowered costs, less chemical use, and a smaller environmental impact. RL Core Technologies is a start-up co-founded by two Amii researchers: Canada CIFAR AI Chairs Martha White and Adam White. With an approach that combines reinforcement learning with the expertise of human operators, RL Core allows control systems to make use of historical data and current conditions, enabling facilities to make adjustments in real time.

The process has seen efficiency gains of up to 10-20% each year, resulting in thousands of dollars of savings in resource costs and equipment maintenance, while making the treatment plants more sustainable. RL Core is currently running three active deployments of its technology in water and wastewater plants in Alberta and British Columbia.

Amii has partnered with pipikwan pêhtâkwan on wâsikan kisewâtisiwin, an AI-powered solution that uses advanced natural language processing (NLP) to combat online hate against Indigenous communities. Using a culturally-sensitive approach led by Indigenous experts, this method can help identify veiled meanings and implications that might otherwise be missed by traditional moderation. 

The solution includes two approaches. First, an AI-enabled plug-in that identifies both overt and covert racist speech on the social media of Indigenous users, automatically hiding harmful comments and making it easier to report.  The use of machine learning allows the tool to more accurately spot coded language that is meant to slip by filters and moderators. 

The second part of the project involves a similar plug-in for non-Indigenous users, which can flag for biased language in writing and make suggestions on how to make language more culturally appropriate. 

In 2024, wâsikan kisewâtisiwin was chosen as a finalist in MIT’s global Solve competition, which identifies projects that are solving critical problems for the public good.

Machine learning is becoming an essential tool in space science, enabling us to learn more about our universe. From analyzing the massive amounts of data collected from satellites and spacecraft to creating autonomous systems that can explore areas far from Earth, artificial intelligence is changing scientific discovery.

Amii Fellow and Canada CIFAR AI Chair Abby Azari’s is also a Science Team Member of the NASA MAVEN mission to Mars where she leads machine learning research relevant to studying Mars’ space environment. Her research applies machine learning to solve challenges in planetary science and space physics and her interdisciplinary contributions have led to further insights into topics like the atmosphere of Mars and Saturn’s magnetosphere.

Complying with regulations is a significant task for many companies, especially those working in high-regulatory industries like energy, utilities, and manufacturing. And it's a burden that continues to grow  - the number of federal regulatory requirements in Canada grew by about 2.1% annually between 2006 and 2021. 

Ultimarii is working with Amii to use AI to change how organizations handle their regulatory workflows. Their platform is trained on thousands of pages of regulatory filings, allowing the models to make insights, automate complex tasks, and flag potential risks before they become an issue. The result is a streamlined regulatory workflow that reduces costs and prevents long delays in getting projects off the ground. Ultimarii recently secured $5.2M USD to scale their platform, intending to support major projects in sectors like energy, mining, and utilities. 

Artificial intelligence is being used to revolutionize the field of medical imaging. Amii Fellow and Canada CIFAR Chair Jacob Jaremko’s work on using machine learning in ultrasounds has made great advances in making ultrasound easier and more accessible to everyone. 

Ultrasound imaging is an extremely common and helpful diagnostic tool. But properly using and interpreting ultrasound images is a specialized skill, requiring extensive training and experience. That has led to a severe shortage of ultrasound technicians in Canada: surveys show that 64% of sonographers express feelings of burnout due to their workload. This also means that many remote and rural communities, especially Indigenous communities, don’t have easy access to ultrasound imaging and must travel long distances for appointments. 

Jaremko’s work has developed machine learning models that can be used with small, portable ultrasound machines, allowing them to be used with less extensive training. The software can guide the user on how to position the scanner to achieve clear images, which are then uploaded into an AI tool trained on thousands of other images to spot anomalies and concerns. It’s not a replacement for trained ultrasound technicians, but Jaremko says it could be a powerful tool for streamlining cases, reducing the workload for human experts, and drastically reducing wait times. One of the research team’s primary focuses is on using the technology to more effectively detect hip dysplasia in infants, which can lead to earlier treatments for a condition that would otherwise be missed. 

This research led to the creation of a successful start-up, Medo.Ai, which focused on commercialization. In 2022, Medo was acquired by ultrasound hardware maker Exo, which is integrating the AI tool into its ultrasound devices.

The University of Alberta’s BLINC Lab (Bionic Limbs for Improved Natural Control), supported by Amii, is transforming research into artificial limbs, leading to new prosthetics that are easier to control and more responsive to the needs of those with limb differences. 


Co-led by Jacqueline Hebert, MD, and Amii Fellow & CIFAR AI Chair Patrick Pilarski, the interdisciplinary lab explores how to improve sensory motor control and integration of advanced prosthetic and robotic systems. By integrating AI approaches like reinforcement learning in the limb control systems, the prosthetics can learn from the individual needs and movements of the person wearing it, becoming more effective through interactions with the user.

The lab’s development team recently received international attention when they took their AI-powered prosthesis to Cybathlon, the premier international competition for assistive technologies, where they placed in the top eight globally in the ARM category.