You have marvelled at the speed of Google search engines, at self-teaching robots.  You’ve been dumbfounded when, on your first visit to a medical specialist, she or she casually pulls up a program containing your entire medical history from the past 10 years.

“Machine learning” uses big data and the ability of today’s computers to sort through and arrange billions upon billions of “data sets” through mathematical formulas. The processes are also referred to as artificial intelligence or, more colloquially, Big Data. 

The ability of massive computer power, harnessed by brilliant machine-learning scientists, is driving the 21st Century. The latest wave of global corporations– Google, Microsoft, Facebook, Lockheed Martin, Yahoo and that great survivor, IBM are all on the leading edge of machine learning.

Edmonton doesn’t have a Google or a Yahoo in our midst – the closest might be global video game producer Bioware, now a division of an even bigger video game maker, Electronic Arts.

But in the heart of the University of Alberta, in its Computing Science Centre, sits one of the top big-data research institutions in North America. 

The Alberta Innovates Centre for Machine Learning is up there with the U of A’s other globally renowned centres such as the Li Ka Shing Institute of Virology and Alberta Diabetes Institute. But outside of its academic and industry circles, the machine-learning centre is low profile. 

Among its members is Dr. Rich Sutton, the inventor and still one of the world’s experts on “reinforcement learning” where robots and ultra-sophisticated computers literally teach themselves through experience. 

Renowned biostatistics researcher Dr. Yutaka Yasui is on a mission to revolutionize disease treatment through individual genetic analysis. Your cancer treatment, for instance, would be based both on conventional diagnostic tools and your own unique DNA traits.

Dr. Michael Bowling’s University of Alberta Computer Poker Research Group was all over the media last January. The team made headlines when its computer program Cepheus beat the world’s best poker players at Heads Up Texas Hold’em poker. Science Magazine then declared the latest version of Cepheus to be unbeatable (at this poker game) by any human. Cepheus achieved this goal with no human expert help, given only the rules of the game … and playing billions and billions of poker games with itself. (You can play poker with Cepheus online at http://poker.srv.ualberta.ca. Good luck.)

Google raised eyebrows world-wide when it paid $400 million for United Kingdom-based Deep Mind, a 50-employee cutting-edge machine-learning research company. Five of Deep Mind’s employees were alumni of Alberta Innovates Centre for Machine Learning research teams.

There are ultra-practical applications of the knowledge pouring out of the machine-learning centre. Dr. Russ Greiner’s medical applications sorts through reams and reams of past patient data to come up with suggested diagnoses and treatments that no one doctor or doctors could do on their own.

Equally interesting is Dr. Griener’s “survival analysis”. Based on computer analysis of millions of past electronic health records plus the patient’s own circumstances, the individual’s likelihood and length of survival after treatment can be more accurately forecast than ever before. 

The accomplishments of the Alberta Innovates Centre for Machine Learning are all the more remarkable given its very small $2 million annual core operating budget – mostly from provincially funded Alberta Innovates Technology Futures.  

Paradox flourishes in this hot-house environment. The University of Alberta centre gives its various team members, mostly University of Alberta students working on their PhDs, invaluable expertise.

Highly educated machine-learning scientists are in high demand around the world. Team members are often offered excellent summer jobs at machine-learning clusters all over the world, followed up by career offers hard to resist.

So how do we keep ’em in Alberta – where our practical scientific expertise centres on the energy sector - to build the province’s fledgling machine-learning commercial sector?

On the optimistic side, the expertise and depth of knowledge built through the Alberta Innovates Centre for Machine Learning has spawned startup machine-learning based companies in the province.

Hopefully, one day, with the right supportive business/intellectual ecosystem, these companies will grow into a home-turf machine-learning commercial cluster. 

FACTOIDS

Highlights of Alberta Innovates Centre for Machine Learning

Richard Sutton co-authors seminal textbook, “Reinforcement Learning: An Introduction” 1998.

Current U of A Dean of Science Jonathan Schaeffer and his team create a checkers-playing computer program unbeatable by humans, 2007.

Prof. Michael Bowling’s Cepheus computer program declared unbeatable at Texas Hold’em poker variation, 2015.

Current collaborations with Industry.

Rail car scheduling optimization

Healthcare

Staffing optimization

Epidemiological/outbreak forecasting

Financial asset allocation and portfolio rebalancing

Data analysis for environmental impact 

Analytics to facilitate and enable change management in organizations

Advanced (medical) Visualization

Advanced prosthesis design

Bio-sample storage and tracking

Brain tumour analysis

Computational game theory

E.coli testing and detection

Diabetes management

Cancer research

Web-based social network analysis and visualization

Graham Hicks

780-707-6379

graham.hicks@hicksbiz.com

www.hicksbiz.com

@hicksbiz