Cory Janssen is Co-Founder and CEO of AltaML. Cory is active in the Alberta entrepreneurial community as a member of the A100 network, the ThresholdImpact University of Alberta Venture Mentoring Service, and as an associate for Creative Destruction Labs Rockies. Cory is also a director at Edmonton Global, is one of three community leaders behind Edmonton.AI.

Until recently, awareness of Alberta’s pre-eminence in the field of artificial intelligence (AI) and particularly machine learning (ML) was largely limited to academic circles, but interest in this “new electricity” is exploding now that we encounter AI/ML-based services routinely (e.g. shopping on Amazon).

The global AI/ML race is on and Alberta is uniquely positioned to succeed, and not just when it comes to startups. Businesses across all industries have the opportunity to develop new sustainable competitive advantages by working with world-class expertise located right here in Alberta.

Let’s start with some basic definitions.

  • Artificial intelligence is a broad umbrella term referring to the pursuit of computer simulation of human intelligence, and research is ongoing in multiple areas.
  • Machine learning is a subarea of AI whereby computers can adapt without being explicitly programmed.
  • Deep learning is a subfield of ML whereby computers can learn representations from data throughout successive layers of increasingly meaningful representations, usually via models called neural networks.

A requirement for ML is data, so there are buzz words around that, too. Data science is a field that overlaps with AI/ML, and involves methods to extract useful information from data, which could be structured (as a spreadsheet or database) or unstructured (such as text, images, video, audio). Big data generally requires massively parallel servers to process large sets of data which is usually continually produced with low information density.

The rise of artificial intelligence in Alberta

Let’s get back to Alberta’s AI advantage. This did not happen overnight. Rather, investment in basic research at the University of Alberta attracted world-class researchers who advanced the field over decades. A major milestone was the establishment of a research lab known as AICML (the Alberta Innovates Centre for Machine Learning) with funding from the Government of Alberta – reflecting their foresight about its future importance. Strategic hires, the most famous of which is Dr. Richard Sutton who literally wrote the textbook on Reinforcement Learning, further cemented the UofA as a world center for AI. AICML was recently spun out of the UofA as part of the Pan-Canadian AI Strategy and is now known as Amii, the Alberta Machine Intelligence Institute. This critical mass of expertise attracted private-sector research infrastructures such as Google’s DeepMind, RBC’s Borealis AI and Mitsubishi Electric. Yes, you read that right: Google is in Edmonton. 

But here’s the issue: we’ve lost a vast majority of talent over the last decade. This is changing now because of two factors. First, we have an emerging cluster of AI startups that are hiring grads, and second, US immigration policy has made it difficult for graduates to go south. So while our AI commercialization lags, we have a unique opportunity right now

This amazing asset was a net exporter of talent for decades and finally, we are turning it around and building an AI ecosystem. What can we do to build on this momentum, and how can forward-thinking Alberta businesses position themselves as early adopters of ML? 

Just as electricity transformed almost everything 100 years ago, today I actually have a hard time thinking of an industry that I don’t think AI will transform in the next several years.

Andrew Ng, computer scientist & Coursera co-founder

Alberta’s opportunity in machine learning

Companies that have embarked upon some form of digital transformation have already taken a crucial step toward ML: data collection. Even if the data is not optimized for ML, it is a potential asset that can be exploited. The question, then, is, how to begin? Our experience is that it’s better to jump in and start a number of small ML projects than embarking on a multi-year journey to perfect your data strategy. The reason being is that there is an interplay between ML and data, and ML will inform how best to adjust your data collection practices – and even what data to collect.

Successful applied ML is really about bringing your subject matter experts together with ML experts, and by focusing on a specific business problem. Machine learning projects will be doomed to failure if data scientists work in silos – they need a well-defined problem with strategic benefit and with a clear line of sight to ROI. The only way you get this is with close collaboration between data science and operations. So whether you build or buy, make sure that the teams are working together. 

And act now. When you look at markets where technology has thrived, be it Israel, Singapore, or Austin, there is connectivity between the large corporates, academia, and the startup ecosystem. We have all the ingredients to transform oil and gas, agriculture, and all our core industries with AI, but we are missing that last link to commercialization. BCA can play a leading role here in building bridges so that the investments in AI research over the last 20 years produce tangible benefits for all Albertans today. 

Share This