“What can machine learning really do?”
That's a question that Walter Schwabe hears a lot. And the CEO of Edmonton's StreamML doesn't hesitate to share his thoughts on the potential for machine learning to revolutionize industries such as healthcare.
"Artificial intelligence will extend the life of the human population on the planet. That's a big statement, but it already is. And in fact, our algorithm has already helped to do that," he says.
"So I see it as something that is going to absolutely continue and permeate the entire health industry at every level."
The second question is usually along the same lines: what can my company do with machine learning?
Schwabe says there's intense interest from industry in artificial intelligence (AI), but also hesitation. Taking the first step into AI can be intimidating, and organizations are unsure of how much they are willing to invest in setting up their own infrastructure.
StreamML aims to simplify that process, providing tools for those interested in learning about machine learning but don't have in-house expertise and resources.
The company got its start in hardware, developing cameras used to gather visual information for machine learning models. But it soon became apparent to StreamML that something was missing. There were organizations out there that wanted to incorporate machine learning into their work but weren't yet ready to invest in building machine learning models from scratch.
StreamML has developed a platform to break down some of those barriers, offered as a software-as-a-service model. First, users sign up for access to StreamML's platform, which specializes in handling visual data. There, they can upload their own data, tag it, and build effective machine learning models without having to code or develop an algorithm of their own.
Making machine learning simpler and faster
"We wanted to make it simpler. We wanted to make it faster. Now, we can help other organizations build their machine learning applications quicker," says StreamML co-founder John Murphy.
For those clients who need something more complex, StreamML can work with them to create custom machine learning models. So far, the company has contributed to projects of wildly different scopes: everything from sorting LEGO pieces by shape, to building an early detection maintenance system for nuclear power plants.
While Schwabe and Murphy both see machine learning playing a big part in the future of everything from agriculture to environmental protection, one industry is of particular interest: healthcare. Schwabe says medical machine learning has the potential to transform the way we deal with our health.
Medicine is a data-heavy industry, Schwabe explains. And machine learning can work with vast amounts of information with speed and accuracy impossible for human beings alone. As a result, medical AI can be an invaluable and versatile tool to assist healthcare providers, leading to better outcomes for patients.
The future of medical AI
The desire to unlock this potential led StreamML to participate in the Ecosystems Project funded by Prairies Canada, to support Alberta organizations to find innovative uses for AI and machine learning in health and wellness. The arrangement was mutually beneficial: companies taking part used StreamML's platform to create models much more quickly and easily than if they had to do so alone. Meanwhile, StreamML was able to use their feedback to refine and expand their own infrastructure.
The company's platform supported the work of organizations like CardiAI, which is investigating whether artificial intelligence could help spot warning signs of heart disease. The model could alert cardiologists that a patient might be at risk for a heart condition. Using machine learning to handle time-consuming tasks like data analysis, physicians can focus their time on the patients at risk, Schwabe says.
"You've got a shortage of professionals in that area. There's always more need than there are people to meet the need. This is where our machine learning algorithm can help," says Schwabe.
Another company involved in the ecosystems project is Bio-Stream Diagnostics, which spun off from StreamML in 2020. Bio-stream is using machine learning to develop a new generation of COVID testing with the potential to be faster and more accurate than current rapid tests.
Schwabe says Alberta has a chance to play a significant role in the future of medical AI. The province already has a long history of being a centre for healthcare innovation. But he says that many aren't aware of how robust the province's machine-learning community is, particularly in Edmonton. It has what he calls a "healthy dollop" of competition and cooperation – like the ecosystems project – that attracts AI talent from around the world.
"People may not know that we have an incredibly deep and robust artificial intelligence community in Edmonton and Alberta. I love the ingenuity and the 'get it done' attitude. That spirit has been immersed in the artificial intelligence community at large," he says.
"You've got a really, really powerful machine literally moving down the road. And for me, it's just exciting to be a part of that."
This project was part of the Western Economic Diversification Canada Regional Innovation Ecosystems (RIE) program. The initiative brought together nine organizations from non-profit, business and academia to establish viable uses for artificial intelligence and machine learning in health and data analytics.