Computer Science Students Take Second Prize in Thales Competition Tackling the Challenge of “Fake News”
By Connie Jeske Crane
Fake news. We intuitively understand what it is. We also know it can spread like wildfire and have hellish real-life consequences – potentially harming private lives, businesses and even democratic elections. What’s maybe not so clear, is how we should fight back.
Recently, at Thales Canada’s 2018 Student Innovation Championship, this thorny issue was put front and centre to student teams. Each year, the Thales competition, which was launched in 2017, asks university students to tackle a different industry problem. The 2018 version saw 52 teams – including one from Ryerson – participating from across the country. Over a period of six weeks, each team was asked to use artificial intelligence (AI) to develop an innovative solution to the challenge of misinformation in online news.
“For the first round, they asked us to have both a slide and a video presentation to explain our solution,” says Kavyan Tirdad, a PhD student in Computer Science and member of the Ryerson team. Supervised by Dr. Alireza Sadeghian, the team also included other computer science students Alex Dela Cruz, a first-year PhD student and Cory Austin, a third-year undergraduate.
At the first cut, the Ryerson team landed among nine finalists selected to travel to Montreal for the final stage of competition, and present before a panel of judges. “We were very proud,” says Tirdad. Besides the excitement of heading to a city that has become one of the world’s leading AI ecosystems, the Ryerson students also got to spend time at Thales’ Centre of Research and Technology for AI eXpertise (cortAIx) and rub shoulders with industry heavyweights in workshops featuring speakers from Thales, event sponsor IVADO, and Microsoft.
Ultimately the Ryerson team achieved an outstanding finish, winning second place for its project, “Opinion and Evidence Analysis,” as well as the $10,000 prize, at the final ceremony attended by numerous Canadian government officials and industry leaders. Asked about their winning solution, Tirdad says, “the motivation of our approach stems from prior experience we gained through our previous deep learning and natural language processing projects.” He describes the Ryerson solution as taking information from a couple of sources and, through a series of steps, transforming this information into a set of evidence, with items labelled either “valid” or “invalid” based on data provided by users on the same topic. “This way we are able to validate the news or article provided by a user as fake or real based on all the other evidence the system acquired from different sources.”
Tirdad easily links this real-world achievement to the comprehensive AI program offered by Ryerson’s Department of Computer Science. “The department offers a large selection of AI courses… artificial intelligence, machine learning, intelligent systems, and deep learning – and that greatly helped us with the fundamental knowledge required in the competition.” He adds that all team members are members of Ryerson’s Computational Intelligence Initiative Lab (CI2). “Under the supervision of Dr. Sadeghian, CI2 lab members conduct fundamental and applied AI research to discover, develop and evaluate algorithms for medical and industrial challenges. Our lab has research collaborations with St. Michael’s Hospital (through the iBEST) and with York University through The Ontario Research Fund – Research Excellence (ORF-RE)/BRAIN Alliance, SOSCIP and a number of partner universities in Italy and Japan.”
As for the prize money, Tirdad says the team members allocated the prize money equally and will use it to keep pressing forward with their work. “We will use the award money to enhance our knowledge in machine learning by attending various conferences, buying more hardware, and learning materials.”