When it comes to the Alberta economy, oil or oil sands is the first thing that comes to mind for many people. But the province is also a leading force in a rising field: artificial intelligence (AI).
The New York Times and Wall Street Journal recently featured Kory Mathewson, a Ph.D. candidate in computer science at the University of Alberta. Mathewson is an accomplished improvisational performance theatre artist. And if you're wondering how that relates to AI, you should know that he performs on stage with a famous partner - an improvisational robot.
Under the guidance of professor Patrick Pilarski of the U of A computer science department, Mathewson has been collaborating with Dr. Piotr Mirowski in London on a human-machine interaction, specifically on how human theatre artists interact with dialog systems (like chatbots) on stage during live improvised theatre.
Mathewson explained that he uses “techniques from deep learning, natural language understanding, automated speech recognition and voice synthesis. These techniques are combined with methods in reinforcement learning and optimization to improve performance on the task of generating imagined worlds for improvised storytelling. I use these methods, deployed on robots, to perform theatre.”
To the uninitiated, artificial intelligence has at least two popular fields at the moment: machine learning and deep learning. For machine learning, the computer is being trained to make predictions based on a vast trove of data. Netflix, for example, uses machine learning to come up with customized movie recommendations for subscribers.
Deep learning, on the other hand, is machine learning on steroids. The computer, in this case, comes up with predictions based on data but also tries to learn some new connections and insights as situations arise. Imagine Amazon’s Alexa calling a restaurant to make a reservation; she interacts, improvises and converses like a real person with a customer service person on the other line. Alexa learns and continues to improve her language skills by continually being exposed to massive amounts of voice data.
Advances in AI are spreading rapidly and impacting almost every industry throughout the world. AI pioneer and Stanford Professor Andrew Ng refers to it as the "new electricity," a technology that will transform every facet of our lives.
"Current technologies in machine learning reduce the marginal cost of making predictions," Mathewson said. "This reduced cost can impact many sectors, industries and our day-to-day lives as humans.”
Mathewson noted that currently, the field is dominated by computing science, so there is still a long way to go before it impacts every industry and every facet of our lives. Even so, the spread of AI innovations continues unabated. Every day, news coverage features some breakthroughs in AI-based computers being able to accurately diagnose diseases or being able to predict - with a high degree of accuracy - the next person who is most likely to default on a loan. The prediction is based on a multitude of data points, including social media postings, cell phone use patterns and many other details.
A little more than five years ago, artificial intelligence was a lonely field being pursued by Professor Geoffrey Hinton at the University of Toronto and some professors at the University of Alberta computer science department. With the availability of big data, the technology all of a sudden became possible and its popularity catapulted on a massive scale, turning Canadian AI researchers into a global sensation.
Dr. Hinton, for instance, now serves as a consultant to Google. He and the computing science department at the U of A (with the likes of Dr. Patrick Pilarski, Dr. Vadim Bulitko, Dr. Rich Sutton, and Dr. Pierre Boulanger), have become beneficiaries of millions of dollars in funding from the Canadian government, which is aiming to boost the nation’s global dominance in AI.
Many Canadian professors who are involved in AI research are also now actively collaborating with Google and other Silicon Valley tech giants to create further advances in the field. As a result, Alberta’s university students have also benefited from Canada’s rise into AI prominence. Mathewson, for example, has obtained funding from the federal government and is now in Europe trying to finish his dissertation research for his doctorate from the Alberta Machine Learning Institute at the University of Alberta. To date, he has already completed a couple of prestigious research stints at Google and Apple. He also currently serves as a machine learning lab scientist at the University of Toronto’s Creative Destruction Lab.
When asked how leaders can further promote and enhance the use of this technology in Alberta, Mathewson says he believes that investing in “Canadian artificial intelligence organizations can help attract hiring of top students and researchers and allow these groups to connect with the tools and infrastructure they need.” He added that “by investing in Alberta-based companies, research groups and university labs, we diversify the resources that support the Alberta economy. This diversification could lead to better job opportunities in Canada, allowing individuals such as myself to stay in cities like Edmonton.”
Investing in Canadian artificial intelligence organizations can help attract hiring of top students and researchers and allow these groups to connect with the tools and infrastructure they need.
To be a global leader in high-paying jobs in the tech sector, Mathewson believes that “continued investment in the universities in Alberta is critical to continue to foster and develop the leaders, innovators and entrepreneurs of the future. Universities encourage students to imagine and build the future.”
Mathewson noted that universities flourish when they are connecting researchers from multiple domains. “My research connects thought leaders from computing science, theatre, psychology, linguistics, engineering, medicine. Without top teachers, lecturers, and researchers, and without intellectual spaces for these minds to connect, conflict and construct the future, Alberta's innovation might be stifled,” he concluded.
So, how does the province go about producing the next generation of global innovators? The answer may lie in the educational pipeline and as well as in personal support systems like community and family.
“Early in my life I was given a computer, a connection to the internet, and training on how to build electronic, mechanical, and software 'things,'" Mathewson said. "I realized how my learning could be sped up through interactions with CD-ROMs of encyclopedias and MIRC chatrooms of my topics of interest. The computer improved my reading, my writing, my social skills, and allowed me to create programs and internet content (remember blogs?). I am always grateful that my father gave me the technological components, and the challenge of putting it together, making it work, and fixing it when it breaks."