Search
  • Profile
  • Artificial Intelligence

AI pioneer Joelle Pineau is transforming personalized health care

by Krista Davidson Sep 28 / 20

joelle- 640x340

Joelle Pineau is using AI to develop a range of smart health care systems. These include robots that care for the elderly, and the SmartWheeler system, which enables people with mobility impairments to navigate easily and safely through crowded environments. 

The Canada CIFAR AI Chair and advisor to CIFAR’s Learning in Machines & Brains program divides her time between Mila and McGill University. Three years ago, she opened Canada’s first Facebook AI Research Lab and is now managing director for Facebook AI Research (FAIR) Global, overseeing labs in Montreal, Seattle, Pittsburgh, and Menlo Park (California). 

Pineau is tirelessly dedicated to using AI to solve problems important to all. She sees health as a significant opportunity to make a difference.

“What I find most compelling about AI for health was that we could potentially use algorithms to solve real-world problems,” she explains.

"Joelle Pineau is a natural leader, one who does not compromise on her principles and scientific rigor, caring for others and our scientific mission, setting an example not just for her highly successful research group but also for Mila, Montreal's AI ecosystem, and the rest of the machine learning community,” says Yoshua Bengio, scientific director of Mila and professor at the Université de Montréal. Bengio is a Canada CIFAR AI Chair and a co-director for the CIFAR Learning in Machines & Brains program.

AI for personalized health care

Pineau recently collaborated with researchers at the Montréal Neurological Institute to improve neuro stimulation, a treatment that delivers electrical stimulation through the brain to prevent seizures in epilepsy patients. 

Using reinforcement learning, doctors can accurately target when and how the stimulation is delivered. The algorithms used are outperforming traditional techniques in animal models of epilepsy and have applications for other diseases, such as cancer.

This research is part of a larger program focused on accelerating the discovery of personalized treatment strategies. By accessing the rich data available through the Canadian health care system, including clinical, genetic and imaging data, Pineau believes we can predict outcomes such as the likelihood of a disease recurring and a patient’s response to a treatment. 

“With health care in particular, there’s a sense I’m solving problems that are important to humanity,” she says.

The nature of Canada’s publicly funded health care system means health data from diverse populations have been captured which could provide AI researchers with rich, relevant and representative data for which to develop robust algorithms. Responsible access to this data could elevate the role of AI in predicting, diagnosing and preventing disease. As part of the CIFAR Pan-Canadian AI Strategy, Pineau and other researchers specializing in AI for health are leading the charge on personalized medicine. 

In July 2020, CIFAR announced the publication of Building a Health Learning System: Report of the Task Force on Artificial Intelligence for Health (AI4H). The report publishes a set of recommendations for an integrated and collaborative approach that could reduce health care costs, improve patient outcomes and catalyze more Canadian-made AI tools and services.

An advocate for diversity and fairness in AI

Pineau’s early experiences as a researcher has prompted her to become a vocal advocate for responsible AI that benefits all. 

One particular area she’s concerned with is the diversity of AI researchers in the development of new AI systems.

Pineau recounts an experience as an early-career researcher when her team was building speech recognition systems for helicopters.

“We were only testing the voices of male speakers and at some point I asked when we would test female voices. It started out as a big gap in the research methodology but the fact that I was on the team prompted us to ask questions,” she explains.

“It’s one of the reasons I spend a lot of time thinking about diversity and inclusion,” she says. “As we develop the tools of AI, we need to do it in a way that is respectful of the full diversity of population, and essentially everyone on the planet.”