AccelerateAB 2018 took place in Edmonton on April 24. The sold-out conference explored the theme of artificial intelligence and machine learning. As A100 Executive Director Cynthia van Sundert said in her message to attendees:
“Our province is home to some of the world’s leading thinking, research, and startups around A.I. It is fortuitous that this year’s event is being held in Edmonton, where the world famous Alberta Machine Learning Institute (Amii) – the global academic leader in A.I. – is located within the University of Alberta.”
The annual conference alternates between Edmonton and Calgary, and always draws an interesting mix of leaders, investors, influencers, and entrepreneurs. The 450-ish in attendance at the Shaw Conference Centre this year were treated to an opening keynote from Scott Penberthy, Google’s Director of Applied AI, a series of AI-related panels, and a closing keynote with Dr. Richard Sutton, a Research Scientists with DeepMind at the University of Alberta.
EEDC’s Cheryll Watson brought opening remarks, and encouraged everyone to “think of ways for Edmonton and Calgary to work together.” She spoke about having “an Alberta mindset” to be more competitive globally than just thinking about the two cities independently.
Rise of Software 2.0
Scott Penberthy opened his talk with some personal history, telling us how he was inspired by Marvin Minsky and Richard Stallman. It wasn’t long though until he was talking about scalars, vectors, tensors, matrices, and more. It was a bit technical for some in the audience I’m sure, but it served as a nice setup to the central premise of his talk: what if you could do math on thought?
“Artificial intelligence is an over-hyped but under-appreciated change,” he told the crowd. Noting that computing power has improved by a billion-fold since 1958, Penberthy highlighted some of the key advances in AI in recent years, including beating humans at image recognition in 2012 and winning at Go in 2016. He cited Ray Kurzweil’s Law of Accelerating Returns to help illustrate why further advances in AI will come more rapidly (and mentioned this video about a rock). Penberthy talked about AutoML, which is basically AI creating AI, as one such advance. He also showed some charts from the popular and accessible Wait But Why series on AI to really drive home the possibilities before us.
“AI is going to be like having a database,” he said. “Soon every business will have this.” Penberthy devoted a good chunk of his time to explaining how AI will help businesses to be more competitive. He touched on TensorFlow, Google’s open source library for dataflow programming, and encouraged the audience to just start playing with the great tools that are now available. “Don’t build it from scratch, leverage existing AI technologies so you can focus on your differentiation,” he said. Kaggle is a good place to start, he suggested.
I appreciated Penberthy’s overview of AI, and also that he called out the success we’ve had here in Alberta in the field. “Canada saw it when no one else saw it.”
Key Takeaways: Panel Sessions
There were two AI-themed panels throughout the day: The Fundamentals of Artificial Intelligence and An Entrepreneurial Journey with Artificial Intelligence. Here are some key takeaways I noted:
- AI is a very broad term, and machine learning is just one technique.
- AI is at the nexus of many disciplines and fields, and provides us with techniques for intelligently making decisions using data.
- Machine learning is useful when it is impossible to hard code a decision or when things are constantly changing, like in the real world.
- Machine learning will find patterns in your data, but you can’t make inferences about the data that you don’t have.
- You need to know why the data is important for the decision you’re trying to make.
- Successful AI projects need both a well-defined problem and data in a usable format.
- Opportunities include biomedical advances and personalized medicine, automation of boring, repetitive, and dangerous jobs, and ways for humans and machines to work better together.
- The reason we have so much opportunity is because of the hardware – we simply didn’t have the power in the past.
- Edmonton is the best place in the world to start a machine learning company!
In the afternoon, the always popular Scaling Eff-Ups panel took place. Some key takeaways from that session:
- Building a business is one of the most intense things you can do.
- If you’re not making mistakes then you’re not reaching high enough.
- Every industry is surrounded by patents so you need to do your homework.
- Pay attention to the people around you and take action, but know that loyalty can be good too.
Much of the insight from the final panel could actually be boiled down to team selection. Many of the “eff-ups” the entrepreneurs discussed involved a member of the team who was selected hastily or who otherwise was not a good fit for the business.
The afternoon featured the startup pitch competition, a staple at the conference. There were pitches from 8 startups this year:
- Fitset: “Experience fitness freedom with easy access to just about every studio & gym in Edmonton with Fitset.”
- IronSight: “A service-hailing technology that strengthens the link between B2C through data-driven dispatching.”
- MicroMech: “Redefining the auto service industry by sending auto mechanics directly to a customer’s door.”
- Mikata Health: “Built a system sing machine learning that helps doctors and their administration teams to eliminate 1-2 hours of data entry each day.”
- Paytickr: “A cloud-based service for small business that has combined time tracking and payroll distribution services into one platform.”
- Skillpics: “A rich networking community where students can showcase their experience, portfolios, resumes and skills to potential employers.”
- Symend: “A FinTech company that is successfully implementing recovery strategy by combining workflow and campaign automation with proven approaches in behavioural science.”
- ShareSmart: “Mitigating costly healthcare data breaches with a system that allows healthcare professionals to take and share patient information securely.”
I previously wrote about Dr. Richard Sutton’s closing keynote on how Edmonton is a world leader in the science of artificial intelligence. You can also check out this Twitter thread on his talk fro Alex Kearney.
Our city’s leadership position in the field (and indeed, Alberta’s) is a key strength that we need to be proud of, and to leverage. It was great to see AccelerateAB shine a light on this. You can find out more about Edmonton’s AI pedigree at Edmonton.AI, a community-driven group with the goal of creating 100 AI and ML companies and projects.
AccelerateAB 2019 will be taking place in Calgary. Follow them on Twitter for updates!