Times & Transcript: UdeM [Université de Moncton] partners with developer on AI project
A small team of computer scientists from l’Université de Moncton is partnering with a Newfoundland software company to develop a knowledge-sharing platform for major companies using artificial intelligence.
The platform, launched by totaliQ, aims to create a software that companies like Cahill Group - one of Canada’s largest construction companies that has agreed to test the product when it’s finished - can use to share knowledge and expertise among employees. Through Mitacs, a non-profit organization that connects private industry with Canadian academia, totaliQ developed a partnership with computer science professor Moulay Akhloufi and his Perception, Robotics and Intelligent Machines (PRIME) research team, who are responsible for constructing the software’s artificial intelligence.
“The last few years of artificial intelligence development have allowed this work to happen,” Moulay said from his office in the Pavillon Rémi-Rossignol building on campus. “We’ve made progress in this area of research that allows us to create systems that mimic the neural networks of the human brain.”
The software is based on the same kind of software that produces popular applications like Amazon Alexa and Google Home - systems that interact with users by providing them with customized information based on their interests and preferences. This is what Moulay calls “machine learning.”
The software totaliQ is developing will act as a system that can provide users with knowledge and information about their field of work by growing its own database of information about the company’s work.
“Knowledge and expertise is based on experience,” explained Moulay. “For example, there are people at engineering companies that are very knowledgeable about fixing cars. But for new employees who aren’t as knowledgeable about fixing cars, there’s a learning curve and it will take time for them to gain the same understanding as those who are experienced. The idea of this knowledge management system is to develop a database of information that the experienced workers have and share their knowledge through this platform so that it’s accessible.”
Andrew Sinclair, founder and CEO of totaliQ, calls it a “recommender system.”
“It’s the same type of algorithm that Netflix uses to offer you movie and TV show recommendations based on what you’ve watched on their platform,” he explained. “The content in totaliQ is knowledge content, and that can take the form of project lessons, stories from experienced people, Q&A exchanges, and so on. The intention is to build an algorithm that will track the users behaviour within the tool and identify what content is most relevant for them.”
The project, which began in February, is expected to be completed in two years.