The University of Ottawa Faculty of Medicine is pleased to announce the results of the highly competitive 2023 round of the Artificial Intelligence (AI) Seed Funding Program.
Building on a successful inaugural round in 2020 and thanks to the generous support of donors Joseph and Amy Ip, the AI Seed Funding Program aims to support research and training initiatives that will enhance the Faculty’s capacity in the field of medical AI.
The Faculty has awarded $50,000 to support the top five project proposals ($10,000 each) led by the following Principal Investigators:
- Steven Hawken (OHRI)
Using machine learning to predict cognitive decline among long-term care residents.
Being able to anticipate cognitive decline can help long-term care residents and their family caregivers plan for care options and address modifiable factors in a compassionate and timely manner. This project will describe cognitive trajectories of long-term care residents in Ontario and use machine learning methods to develop a tool to predict cognitive decline. - Mathieu Lavallée-Adam (Dept. Biochemistry, Microbiology and Immunology)
RealCoLD: A supervised learning approach for real-time detection of cross-linked peptides using mass spectrometry for immunotherapeutics development
Therapeutic strategies that use the human immune system and supplement it with specific molecules show great potential in cancer and infection treatment. This project aims to use AI to develop software that guides mass spectrometry to gather data that will enable the design of better therapeutics targeting cancer and infections. - Ahmed Nasr (Dept. Surgery, CHEO)
Evaluation of a pediatric laparoscopic simulator using artificial intelligence technology
This project will develop an Artificial Intelligence Pediatric Laparoscopic Simulator (AI PLS) to provide a low-cost, highly accurate training tool for pediatric surgeons. Novice surgeons using the AI PLS will be able to learn surgical techniques specific to operating on children by mirroring the surgical techniques of experts. - Rebecca Thornhill (OHRI)
Better together: developing a patient- and radiologist-informed framework for explainable machine learning
AI programs can now be trained with thousands of medical images to identify the most worrisome cancers. However, we still need ways to explain how AI programs make decisions. One possibility is to use everyday image concepts, such as “wavy” or “blotchy”, to explain this information. This study will invite participants with and without medical backgrounds to identify image concepts that best explain AI results. - Alissa Visram (Dept. Medicine, OHRI)
Optimizing prediction of disease progression from precursor plasma cell disorders to hematologic cancer using artificial intelligence
Currently used risk stratification for patients with precursor plasma cell disorders incorporates limited data and therefore does not accurately predict risk of developing a blood cancer over time. This study will use AI to develop a risk prediction model based on longitudinal clinical and laboratory data that can be applied to provide physicians and patients with real-time, tailored feedback on progression risk.