Heavy investment by government and industry is making Canada an AI powerhouse.
Manufacturers should be paying close attention to developments in artificial intelligence (AI). We’re not talking Skynet, killer robot type technology from the Terminator movies, but rather something that provides endless opportunities to improve efficiency and productivity.
Machine learning is now capable of handling algorithms in ways once thought impossible. Before, algorithms had to be adjusted manually to adapt to changing environments or operations. Today, much smarter machines learn and adjust on their own.
These deep mathematical models, called neural nets, are based on the networks of the human brain, and they enable the analysis of incredible quantities of data to – among other things – handle discrete tasks.
Some of the leading developments in this advanced technology are coming from Canada, which is developing into a centre of excellence for AI research. Central to this excellence is Toronto’s Vector Institute, established through a government-industry partnership. Its mission is to help startups grow next generation technology without leaving Canada; and entice foreign companies to set up their R&D centres here. Last year the province attracted $2.84 billion in investment.
Looking ahead, an IoT/data analysis report from Frost & Sullivan, a San Antonio, Tex.-based consulting firm, concludes development of smart and safe robots using machine learning techniques will be a prime area of focus.
While humans are still needed for logical and reasonable decision making, the report notes cognitive technologies that allow machines to detect changing manufacturing scenarios and then respond in real time will lessen the need for hands-on intervention. But manufacturers need to make significant investments in AI to benefit from these advances.
A number of Canadian industry heavyweights have joined the AI movement, including RBC and Scotiabank, auto parts manufacturer Magna and Canada’s crown investment arm BDC Capital. Together they’ve provided no-strings-attached funding worth $200,000 to startups as part of the NextAI program.
Ride-sharing giant Uber is in with a multi-year pledge, including $5 million to establish a research hub at the Vector Institute to improve autonomous vehicle technology.
The hub will be led by University of Toronto professor and well-known machine perception researcher Raquel Urtasun. Her research has focused on developing software that allows self-driving cars to “see” objects around them.
Magna has invested $5 million in the Vector Institute to attract, develop and retain homegrown AI technology and the people behind it.
The auto parts giant’s strategy focuses on two key areas.
In manufacturing, it involves leveraging AI to provide human operators with enhanced information for decision making; automating quality, sorting and material handling to reduce production costs; and applying AI-enhanced predictive maintenance systems to ensure greater machine up-time.
The other area is autonomous driving. As future mobility evolves, AI will play a major role in dynamic decision-making. Object detection and classification combined with scene segmentation, including traffic volume, speeds and road conditions (weather, light, visibility), will be key.
In Ottawa, Raven Telemetry AI has developed a powerful, user-friendly tool meant to encourage manufacturers to leverage AI by easily making sense of big data.
The company’s solution collects data from shop floor machines for real-time analysis and documents the results. Martin Cloake, Raven’s CEO, says the product interprets massive amounts of raw data before sending it to supervisors and operators so they can act on it to accelerate decision-making and problem-solving.
This is in contrast to a more traditional approach, which is to deploy management tools such as Microsoft Excel. It collects performance numbers in enormous files, then presents the data in graphs and charts that require engineering expertise to understand.
Cloake has found many operators don’t use the data because the information doesn’t provide meaningful suggestions for improving operations.
Raven’s AI solution crunches the data to provide a simple direction on which shop floor operators can act. Monitoring processes quickly detects anomalies that could be problems or identifies areas that contribute to improvements. Cloake says Raven’s clients see, on average, increased overall equipment efficiency of 15% over six months, resulting in close to $4 million in additional output per year using existing staff and the same equipment.
In manufacturing, money talks and AI is poised to make the conversation a lot more interesting. Canada playing a leading role in the field will be good for manufacturers who adapt the technology.
Frost & Sulliva
This article appeared in the September 2017 issue of PLANT Magazine.