Hi everyone, 

My name is Eirene Seiradaki, and I’m the Director of Research Partnerships at Borealis AI. We’ve just launched this new website this week, and, since this is our inaugural blog post, I thought I’d take this opportunity to introduce who we are and what we do. 

Borealis AI is the R&D arm of the Royal Bank of Canada, and we perform fundamental research in machine learning theory and applications. Some of our areas of research include: Reinforcement Learning, Natural Language Processing, Deep Learning, Optimization in Machine Learning, and High Performance Computing. 

In September 2016, our team was established in order to address an increasingly urgent – and growing – need in this field. As referenced in a recent Globe and Mail article, “Artificial Intelligence is the future, and Canada can seize it,” there exists a paradox: although Canadian educational institutions have produced some of the world’s most brilliant minds in AI throughout the last ten years, Canada continues to suffer from the ongoing “brain drain” to the states. Leading scholars in the AI frontier at all academic levels, including PhD students, post-doctoral fellows, as well as noted faculty members at prestigious institutions, in a quest for resources and proper support to perform their research, have increasingly been moving south of the border to pursue opportunities. Top talent are finding homes at other universities and even at major corporations and startups in Silicon Valley.

The cause of this problem is well-known: Canada doesn’t have enough world-class research labs by which to retain its world-class AI talent. Furthermore, we know that top talent is motivated by being able to perform world-leading research, and to do this, they need access to datasets. We realized that RBC not only has the meaningful problems to solve – which will positively impact millions of people around the world – but also the datasets required to find ML solutions to such problems. And, thus, RBC Research was forged.

We are a research lab that provides leading scholars in machine learning with the environment, resources, and vision to conduct world-class research. Our pillars say it all:

Academic Freedom

Our researchers are in the driver’s seat. They are not merely handed projects to complete; rather, they are empowered to steer their research wherever they see where it could produce the greatest impact. Our group aims to publish our work in top-tier academic conferences and journals as well as collaborate with top academic groups openly and transparently. Our locations reinforce this very mission: our Toronto office is at the University of Toronto, and our Edmonton office is at the University of Alberta. 

Data

The backbone of every machine learning system is data. We are fortunate to have access to massive datasets from decades of real market data (structured and unstructured), which can be used for NLP, deep learning, reinforcement learning, and others.

Computational Power

Think of the most beautiful car you’ve ever seen. Sleek design, sharp paint job, perfect manufacturing. Now put a golf cart engine in that car. What you have now is all the best parts without the mechanism to go tearing down the highway. Datasets without sufficient computational power are like that car without a proper engine. At Borealis AI, we have the computational resources to match our massive datasets, enabling our researchers to solve the next AI problems with all the right resources.

We aspire to work with people who share our passion for open research, high-quality publications, and all things machine learning. We are excited about where this will take us, and we hope that you will continue to follow this journey with our team. We will be writing regular blog posts to share our activities and research with you, and we hope to have the opportunity to connect with you soon, either online here, in Toronto, Edmonton, or at one of the conferences we will be attending this year. Keep checking back here for all of the latest updates and here’s to a banner 2017! 

Cheers,
Eirene