SASKATCHEWAN — The world is seeing an explosion in the use of artificial intelligence, also known as AI, not just in business and industry but in common daily life. Most people have given it at least a passing thought in recent months—whether considering the benefits or pitfalls of using a generative AI platform such as ChatGPT for work or education or in worrying whether it will get away from us if we don’t set government policy on its development and use.
More often, though, people encounter AI and subsets such as machine learning and natural language processing without even thinking about it, such as when asking Siri or Alexa what the day’s temperature will be with windchill — especially in Saskatchewan in winter. In 2024, AI has become that ubiquitous.
Saskatchewan Polytechnic’s Dr. Terry Peckham, who oversees the province’s first Technology Access Centre (TAC), is exponentially more excited than afraid of AI. As director of the Peckham welcomes applied research projects that explore AI’s application to solve real world problems in a variety of sectors. “Using AI isn’t lazy, as some people think,” says Peckham, “and we’re not at risk of seeing it take over from humans. Like all technology, it’s a tool—a whole bunch of tools, really.”
DICE currently has a half dozen or more projects underway that explore the use of AI. “We’re starting to see an increase in the ways AI can be applied” adds Peckham, “as well as the fine-tuning that will allow it to provide more elegant solutions to complex problems. That’s incredibly exciting!”
One example is gravel processing—an industry that doesn’t typically bring AI, or even computers, to mind. For McIntosh Sand and Gravel Ltd., a Saskatoon company that performs gravel crushing and processing for industry and municipalities throughout Saskatchewan, machine learning and image processing are new tools the company is investigating with the help of DICE researchers. “Our mission is to develop and utilize new technologies that will help decrease costs and streamline processes,” says COO Connor Regier. “In certain scenarios, digital technologies can modernize or complement older methods in an industry that hasn’t changed much in the past fifty years or more.”
Sask Polytech is working with the company to develop a digital method to determine gravel particulate size, a process traditionally done through a sieve analysis that takes significant time and is costly. Research associate Deshan Kanaththage is working with the company to develop a portable system that uses AI to help expedite results, which will reduce analysis cost and allow for quicker decision making.
Data is at the core of everything DICE does, and Peckham explains that, “Where there is data, AI can be of help—whether in sorting and extracting or organizing and analyzing.” Healthcare, a sector with a long history of gathering and storing patient data, is correspondingly a rapidly growing area for AI use. “It’s certainly a space that stands to see significant benefits from the application of artificial intelligence,” says Peckam, “and we’re starting to see interest from applied research partners in this field looking to integrate AI into their systems.”
One such partner is Cardea Health, Inc., an Ontario-based medical technology company working with Sask Polytech researcher Azizi Mehrnoosh to develop an AI-based platform for cardiac specialists. Their goal is a system that can use AI to extract non-standardized patient data from multiple sources into a coherent data flow and single source.
Cardea’s chief technology officer Sriram Gowrishankar, a veteran health care technologist with a passion for data, explains, “Our current project involves employing optical character recognition (OCR) and natural language processing (NLP) technologies to decipher complex medical reports such as electrocardiograms and echocardiograms with the primary goal of identifying critical risk factors for cardiovascular diseases.” The challenge lies in adapting NLP algorithms to the nuances of medical language, particularly within cardiology, where maintaining specialized vocabularies and lexicons is essential for accurate analysis. “This initiative,” says Gowrishankar, “aims to enhance patient care by efficiently pinpointing cardiovascular risk factors, facilitating the potential qualification of at-risk individuals for pertinent clinical trials.”
Agriculture is another industry where the application of AI is rapidly changing standard practices. Canadian grain growers are paid based on Canadian Grain Commission grading factors. Current grading methods that aggregate harvest data are susceptible to inaccuracy when gathered offsite, and inaccuracy ultimately means lower grain prices or the inability to tap into pricing premiums for growers.
AgTech company Ground Truth Agriculture, Inc. is working with Sask Polytech researchers Nathaniel Vallee and Prakher Singh to advance their automated grain grading platform and include the ability to grade red lentils by identifying diseased and damaged red lentil kernels and impurities in the sample. Using machine vision tools the system they are developing will improve the speed and accuracy of grain quality assessment at harvest and all points of the grain supply chain.
“What we’re developing will help growers to market their grains with confidence,” says Ground Truth Agriculture chief operating officer, Divyesh Patel. “We had a vision and DICE’s digital expertise has helped guide us. Having access to DICE’s expertise has meant we have been able to develop an award-winning system faster.”
Another example of process improvement through AI is in a complex mining environment. Based in Saskatchewan, international mining company on a progression of projects exploring the application of AI to make improvements to the water jet boring system (JBS) they use to safely extract uranium from Saskatchewan’s large deposits. With funding from the Natural Science and Engineering Research Council of Canada (NSERC) and the International Minerals Innovation Institute (IMII), Cameco and DICE set out on a journey of discovery that has led to some promising preliminary results.
Cameco and DICE worked diligently over a two-year period to develop AI models to improve predictability of mining recovery and dilution using the JBS mining method. Based on the results of training using historical data, Cameco then commissioned a pilot project in which the AI model, in conjunction with human engineers, mined new cavities to identify real-world improvements. The pilot study indicated potential to achieve improvement in ore production using the AI models, with a reduction in the waste produced (tailings) and a further reduction in energy consumption. The pilot study also indicated other areas for improvement in the model that should further impact overall results.
“This trifecta of improved production with reduced waste and reduced energy consumption is something that both DICE and Cameco are proud of,” says Peckham. “With further opportunities for this research to mature, indications for success are extremely positive. We continue to work together to improve and mature the models and processes.”
As DICE at Sask Polytech begins its tenth year, Peckham and his research team look forward to helping new partners find creative ways to put AI to good use. He and his team are grateful for the ongoing operational TAC funding made possible by Innovation Saskatchewan and the Natural Sciences and Engineering Research Council of Canada (NSERC) that positions the centre to delve further into this growing field of applied research with new clients. “We’ve only seen the beginning of what AI and related digital technologies can do,” says Peckham. “It’s a great privilege to be working alongside our partners in various sectors to explore and put into practice new ways of working with data. We all stand to benefit when new areas of knowledge are put into practical applications in our day-to-day lives.”
— Submitted by Sask Poltytech Media Relations
